id
stringlengths 9
16
| submitter
stringlengths 1
64
⌀ | authors
stringlengths 3
62.8k
| title
stringlengths 1
382
| comments
stringlengths 1
1.15k
⌀ | journal-ref
stringlengths 1
557
⌀ | doi
stringlengths 7
153
⌀ | report-no
stringlengths 1
509
⌀ | categories
stringlengths 5
125
| license
stringclasses 9
values | abstract
stringlengths 6
6.09k
| versions
listlengths 1
187
| update_date
timestamp[s] | authors_parsed
sequencelengths 1
2.83k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2404.11629 | Lei Zhou | Lei Zhou | Rules and Algorithms for Objective Construction of Fuzzy Sets | null | null | null | null | math.GM | http://creativecommons.org/licenses/by/4.0/ | This paper aims to present objective methods for constructing new fuzzy sets
from known fuzzy or classical sets, defined over the elements of a finite
universe's superstructure. The paper proposes rules for assigning membership
functions to these new fuzzy sets, leading to two important findings. Firstly,
the property concerning the cardinality of a power set in classical theory has
been extended to the fuzzy setting, whereby the scalar cardinality of a fuzzy
set $\tilde B$ defined on the power set of a finite universe of a fuzzy set
$\tilde A$ satisfies $\text{card}(\tilde B)=2^{\text{card}(\tilde A)}$.
Secondly, the novel algorithms allow for an arbitrary membership value to be
objectively achieved and represented by a specific binary sequence.
| [
{
"version": "v1",
"created": "Tue, 16 Apr 2024 08:19:11 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 04:37:25 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Zhou",
"Lei",
""
]
] |
2404.12096 | Dawei Zhu | Dawei Zhu, Liang Wang, Nan Yang, Yifan Song, Wenhao Wu, Furu Wei,
Sujian Li | LongEmbed: Extending Embedding Models for Long Context Retrieval | EMNLP 2024 Camera Ready | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | Embedding models play a pivot role in modern NLP applications such as IR and
RAG. While the context limit of LLMs has been pushed beyond 1 million tokens,
embedding models are still confined to a narrow context window not exceeding 8k
tokens, refrained from application scenarios requiring long inputs such as
legal contracts. This paper explores context window extension of existing
embedding models, pushing the limit to 32k without requiring additional
training. First, we examine the performance of current embedding models for
long context retrieval on our newly constructed LongEmbed benchmark. LongEmbed
comprises two synthetic tasks and four carefully chosen real-world tasks,
featuring documents of varying length and dispersed target information.
Benchmarking results underscore huge room for improvement in these models.
Based on this, comprehensive experiments show that training-free context window
extension strategies like position interpolation can effectively extend the
context window of existing embedding models by several folds, regardless of
their original context being 512 or beyond 4k. Furthermore, for models
employing absolute position encoding (APE), we show the possibility of further
fine-tuning to harvest notable performance gains while strictly preserving
original behavior for short inputs. For models using rotary position embedding
(RoPE), significant enhancements are observed when employing RoPE-specific
methods, such as NTK and SelfExtend, indicating RoPE's superiority over APE for
context window extension. To facilitate future research, we release E5-Base-4k
and E5-RoPE-Base, along with the LongEmbed benchmark.
| [
{
"version": "v1",
"created": "Thu, 18 Apr 2024 11:29:23 GMT"
},
{
"version": "v2",
"created": "Thu, 25 Apr 2024 02:26:15 GMT"
},
{
"version": "v3",
"created": "Thu, 7 Nov 2024 09:29:32 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Zhu",
"Dawei",
""
],
[
"Wang",
"Liang",
""
],
[
"Yang",
"Nan",
""
],
[
"Song",
"Yifan",
""
],
[
"Wu",
"Wenhao",
""
],
[
"Wei",
"Furu",
""
],
[
"Li",
"Sujian",
""
]
] |
2404.12669 | Ushnanshu Dutta Dr | Chandrima Mallick, Ushnanshu Dutta, Moumita Bhowmik, Greeshma M.
Mohan, Anupam Hazra, S. D. Pawar, and Jen-Ping Chen | Comparison of Two-Moment and Three-Moment Bulk Microphysics Schemes in
Thunderstorm Simulations over Indian Subcontinent | null | null | 10.1016/j.atmosres.2024.107614 | null | physics.ao-ph physics.data-an | http://creativecommons.org/licenses/by/4.0/ | We have performed three-dimensional thunderstorm real simulations using the
two-moment and three-moment bulk microphysics schemes in the Weather Research
and Forecasting (WRF) model. We have analyzed three cases to understand the
potential differences between the double-moment (Morrison-2M) and National
Taiwan University triple-moment (NTU-3M) microphysics parameterizations in
capturing the characteristics of lightning events over the Indian subcontinent.
Despite general resemblances in these schemes, the simulations reveal distinct
differences in storm structure, cloud hydrometeors formation, and
precipitation. The lightning flash counts from the in situ lightning detection
network (LDN) are also used to compare the simulation of storms. The Lightning
Potential Index (LPI) is computed for Morrison-2M and NTU-3M microphysics
schemes and compared it with the Lightning Detection Network (LDN) observation.
In most cases, the Morrison-2M shows more LPI than the NTU-3M scheme. Both the
schemes also differ in simulating rainfall and other thermodynamical,
dynamical, and microphysical parameters in the model. Here, we have attempted
to identify the basic differences between these two schemes, which may be
responsible for the discrepancies in the simulations. In particular, the
Morrison-2M produced much higher surface precipitation rates. The effects on
the size distributions cloud hydrometeors between two microphysical schemes are
important to simulate the biases in the precipitation and lightning flash
counts. The inclusions of ice crystal shapes are responsible for many of the
key differences between the two microphysics simulations. Different approaches
in treating cloud ice, snow, and graupel may have an impact on the simulation
of lightning and precipitation. Results show that the simulation of lightning
events is sensitive to microphysical parameterization schemes in NWP models.
| [
{
"version": "v1",
"created": "Fri, 19 Apr 2024 07:12:38 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Mallick",
"Chandrima",
""
],
[
"Dutta",
"Ushnanshu",
""
],
[
"Bhowmik",
"Moumita",
""
],
[
"Mohan",
"Greeshma M.",
""
],
[
"Hazra",
"Anupam",
""
],
[
"Pawar",
"S. D.",
""
],
[
"Chen",
"Jen-Ping",
""
]
] |
2404.12962 | George Gkountoumis | George Gkountoumis | Asymmetric $\mathbb{Z}_4$ orbifolds of type IIB string theory revisited | 6 pages + appendix and references. arXiv admin note: text overlap
with arXiv:2302.09112 | null | null | null | hep-th | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We construct freely acting asymmetric $\mathbb{Z}_4$ orbifolds of type IIB
string theory on $T^5$ preserving 24,16 or 8 supercharges in five dimensions.
We show that these models are well-defined if the SO(8) lattice is chosen,
instead of the SU(2)$^4$ lattice, which was previously considered in the
literature.
| [
{
"version": "v1",
"created": "Fri, 19 Apr 2024 15:45:33 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 09:22:24 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Gkountoumis",
"George",
""
]
] |
2404.13329 | Jesse Railo | Jesse Railo | A note on the Fourier magnitude data and Sobolev embeddings | 11 pages, accepted manuscript | J Fourier Anal Appl 30, 59 (2024) | 10.1007/s00041-024-10116-9 | null | math.FA | http://creativecommons.org/licenses/by/4.0/ | We study Sobolev $H^s(\mathbb{R}^n)$, $s \in \mathbb{R}$, stability of the
Fourier phase problem to recover $f$ from the knowledge of $|\hat{f}|$ with an
additional Bessel potential $H^{t,p}(\mathbb{R}^n)$ a priori estimate when $t
\in \mathbb{R}$ and $p \in [1,2]$. These estimates are related to the ones
studied recently by Steinerberger in "On the stability of Fourier phase
retrieval" J. Fourier Anal. Appl., 28(2):29, 2022. While our estimates in
general are different, they share some comparable special cases and the main
improvement given here is that we can remove an additional imaginary term and
obtain sharper constants. We also consider these estimates for the quotient
distances related to the non-uniqueness of the Fourier phase problem. Our
arguments closely follow the Fourier analysis proof of the Sobolev embeddings
for Bessel potential spaces with minor modifications.
| [
{
"version": "v1",
"created": "Sat, 20 Apr 2024 09:08:58 GMT"
},
{
"version": "v2",
"created": "Fri, 20 Sep 2024 16:53:53 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Railo",
"Jesse",
""
]
] |
2404.15036 | Brendan Unikewicz | B. M. Unikewicz, A. M. Pincot, T. Cohen | An Accessible Instrument for Measuring Soft Material Mechanical
Properties | null | null | null | null | physics.ins-det physics.data-an | http://creativecommons.org/licenses/by/4.0/ | Soft material research has seen significant growth in recent years, with
emerging applications in robotics, electronics, and healthcare diagnostics
where understanding material mechanical response is crucial for precision
design. Traditional methods for measuring nonlinear mechanical properties of
soft materials require specially sized samples that are extracted from their
natural environment to be mounted on the testing instrument. This has been
shown to compromise data accuracy and precision in various soft and biological
materials. To overcome this, the Volume Controlled Cavity Expansion (VCCE)
method was developed. This technique tests soft materials by controlling the
formation rate of a liquid cavity inside the materials at the tip of an
injection needle, and simultaneously measuring the resisting pressure which
describes the material response. Despite VCCE's early successes, expansion of
its application beyond academia has been hindered by cost, size, and expertise.
In response to this, the first portable, bench-top instrument utilizing VCCE is
presented here. This device, built with affordable, readily available
components and open-source software, streamlines VCCE experimentation without
sacrificing performance or precision. It is especially suitable for
space-limited settings and designed for use by non-experts, promoting
widespread adoption. The instrument's efficacy was demonstrated through testing
Polydimethylsiloxane (PDMS) samples of varying stiffness. This study not only
validates instrument performance, but also sets the stage for further
advancements and broader applications in soft material testing. All data, along
with acquisition, control, and post-processing scripts, are made available on
GitHub.
| [
{
"version": "v1",
"created": "Tue, 23 Apr 2024 13:41:10 GMT"
},
{
"version": "v2",
"created": "Thu, 19 Sep 2024 22:02:05 GMT"
},
{
"version": "v3",
"created": "Mon, 23 Sep 2024 20:38:50 GMT"
},
{
"version": "v4",
"created": "Wed, 6 Nov 2024 19:42:00 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Unikewicz",
"B. M.",
""
],
[
"Pincot",
"A. M.",
""
],
[
"Cohen",
"T.",
""
]
] |
2404.15050 | Vladimir Mazurenko Dr. | Oleg M. Sotnikov, Ilia A. Iakovlev, Evgeniy O. Kiktenko, Aleksey K.
Fedorov, Vladimir V. Mazurenko | Achieving the volume-law entropy regime with random-sign Dicke states | The second version of the manuscript with new interesting results | null | null | null | quant-ph cond-mat.dis-nn cond-mat.str-el | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Manipulating entanglement, which reflects non-local correlations in a quantum
system and defines the complexity of describing its wave function, represents
the extremely tough challenge in the fields of quantum computing, quantum
information, and condensed matter physics. In this work, by the example of the
well-structured Dicke states we demonstrate that the complexity of these
real-valued wave functions can be accurately tuned by introducing a random-sign
structure, which allows us to explore the regime of the volume-law
entanglement. Importantly, setting nontrivial sign structure one can increase
the entanglement entropy of the Dicke state to the values that are close to
Page's estimates for Haar-random states. The practical realization of these
random-sign Dicke states is possible on different physical platforms with
shallow quantum circuits. On the level of the measurements the change in the
quantum state complexity due to sign structure can be traced out with the
dissimilarity measure that estimates multi-scale variety of patterns in
bit-string arrays.
| [
{
"version": "v1",
"created": "Tue, 23 Apr 2024 13:50:07 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 09:56:12 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Sotnikov",
"Oleg M.",
""
],
[
"Iakovlev",
"Ilia A.",
""
],
[
"Kiktenko",
"Evgeniy O.",
""
],
[
"Fedorov",
"Aleksey K.",
""
],
[
"Mazurenko",
"Vladimir V.",
""
]
] |
2404.15278 | Kongyang Chen | Wenjun Lan, Kongyang Chen, Jiannong Cao, Yikai Li, Ning Li, Qi Chen,
Yuvraj Sahni | Security-Sensitive Task Offloading in Integrated Satellite-Terrestrial
Networks | null | IEEE Transactions on Mobile Computing, 2024 | 10.1109/TMC.2024.3489619 | null | eess.SP cs.CR cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the rapid development of sixth-generation (6G) communication technology,
global communication networks are moving towards the goal of comprehensive and
seamless coverage. In particular, low earth orbit (LEO) satellites have become
a critical component of satellite communication networks. The emergence of LEO
satellites has brought about new computational resources known as the
\textit{LEO satellite edge}, enabling ground users (GU) to offload computing
tasks to the resource-rich LEO satellite edge. However, existing LEO satellite
computational offloading solutions primarily focus on optimizing system
performance, neglecting the potential issue of malicious satellite attacks
during task offloading. In this paper, we propose the deployment of LEO
satellite edge in an integrated satellite-terrestrial networks (ISTN) structure
to support \textit{security-sensitive computing task offloading}. We model the
task allocation and offloading order problem as a joint optimization problem to
minimize task offloading delay, energy consumption, and the number of attacks
while satisfying reliability constraints. To achieve this objective, we model
the task offloading process as a Markov decision process (MDP) and propose a
security-sensitive task offloading strategy optimization algorithm based on
proximal policy optimization (PPO). Experimental results demonstrate that our
algorithm significantly outperforms other benchmark methods in terms of
performance.
| [
{
"version": "v1",
"created": "Sat, 20 Jan 2024 07:29:55 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Lan",
"Wenjun",
""
],
[
"Chen",
"Kongyang",
""
],
[
"Cao",
"Jiannong",
""
],
[
"Li",
"Yikai",
""
],
[
"Li",
"Ning",
""
],
[
"Chen",
"Qi",
""
],
[
"Sahni",
"Yuvraj",
""
]
] |
2404.16783 | Xiehang Yu | Xie-Hang Yu, J. Ignacio Cirac, Pavel Kos, and Georgios Styliaris | Dual-isometric Projected Entangled Pair States | 5+10 pages, 1 figure | 10.1103/PhysRevLett.133.190401 (2024) | null | null | quant-ph cond-mat.stat-mech cond-mat.str-el | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Efficient characterization of higher dimensional many-body physical states
presents significant challenges. In this paper, we propose a new class of
Project Entangled Pair State (PEPS) that incorporates two isometric conditions.
This new class facilitates the efficient calculation of general local
observables and certain two-point correlation functions, which have been
previously shown to be intractable for general PEPS, or PEPS with only a single
isometric constraint. Despite incorporating two isometric conditions, our class
preserves the rich physical structure while enhancing the analytical
capabilities. It features a large set of tunable parameters, with only a
subleading correction compared to that of general PEPS. Furthermore, we
analytically demonstrate that this class can encode universal quantum
computations and can represent a transition from topological to trivial order.
| [
{
"version": "v1",
"created": "Thu, 25 Apr 2024 17:31:31 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Jun 2024 12:10:47 GMT"
},
{
"version": "v3",
"created": "Wed, 25 Sep 2024 14:34:14 GMT"
},
{
"version": "v4",
"created": "Thu, 7 Nov 2024 15:19:31 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Yu",
"Xie-Hang",
""
],
[
"Cirac",
"J. Ignacio",
""
],
[
"Kos",
"Pavel",
""
],
[
"Styliaris",
"Georgios",
""
]
] |
2404.17008 | Arno Botha | Arno Botha, Tanja Verster, Roelinde Bester | The TruEnd-procedure: Treating trailing zero-valued balances in credit
data | 21 pages, 7545 words, 10 Figures | null | null | null | q-fin.RM q-fin.ST stat.AP | http://creativecommons.org/licenses/by/4.0/ | A novel procedure is presented for finding the true but latent endpoints
within the repayment histories of individual loans. The monthly observations
beyond these true endpoints are false, largely due to operational failures that
delay account closure, thereby corrupting some loans in the dataset with
`false' observations. Detecting these false observations is difficult at scale
since each affected loan history might have a different sequence of zero (or
very small) month-end balances that persist towards the end. Identifying these
trails of diminutive balances would require an exact definition of a "small
balance", which can be found using our so-called TruEnd-procedure. We
demonstrate this procedure and isolate the ideal small-balance definition using
residential mortgages from a large South African bank. Evidently, corrupted
loans are remarkably prevalent and have excess histories that are surprisingly
long, which ruin the timing of certain risk events and compromise any
subsequent time-to-event model such as survival analysis. Excess histories can
be discarded using the ideal small-balance definition, which demonstrably
improves the accuracy of both the predicted timing and severity of risk events,
without materially impacting the monetary value of the portfolio. The resulting
estimates of credit losses are lower and less biased, which augurs well for
raising accurate credit impairments under the IFRS 9 accounting standard. Our
work therefore addresses a pernicious data error, which highlights the pivotal
role of data preparation in producing credible forecasts of credit risk.
| [
{
"version": "v1",
"created": "Thu, 25 Apr 2024 20:07:33 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 09:41:26 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Botha",
"Arno",
""
],
[
"Verster",
"Tanja",
""
],
[
"Bester",
"Roelinde",
""
]
] |
2404.17123 | Jianlong Chen | Wei Xu, Jianlong Chen, Zhicheng Ding, Jinyin Wang | Text Sentiment Analysis and Classification Based on Bidirectional Gated
Recurrent Units (GRUs) Model | accepted by the 2nd International Conference on Software Engineering
and Machine Learning (CONF-SEML 2024) | Applied and Computational Engineering. 77 (2024) 132-137 | 10.54254/2755-2721/77/20240670 | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper explores the importance of text sentiment analysis and
classification in the field of natural language processing, and proposes a new
approach to sentiment analysis and classification based on the bidirectional
gated recurrent units (GRUs) model. The study firstly analyses the word cloud
model of the text with six sentiment labels, and then carries out data
preprocessing, including the steps of removing special symbols, punctuation
marks, numbers, stop words and non-alphabetic parts. Subsequently, the data set
is divided into training set and test set, and through model training and
testing, it is found that the accuracy of the validation set is increased from
85% to 93% with training, which is an increase of 8%; at the same time, the
loss value of the validation set decreases from 0.7 to 0.1 and tends to be
stable, and the model is gradually close to the actual value, which can
effectively classify the text emotions. The confusion matrix shows that the
accuracy of the model on the test set reaches 94.8%, the precision is 95.9%,
the recall is 99.1%, and the F1 score is 97.4%, which proves that the model has
good generalisation ability and classification effect. Overall, the study
demonstrated an effective method for text sentiment analysis and classification
with satisfactory results.
| [
{
"version": "v1",
"created": "Fri, 26 Apr 2024 02:40:03 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Jun 2024 14:12:17 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Xu",
"Wei",
""
],
[
"Chen",
"Jianlong",
""
],
[
"Ding",
"Zhicheng",
""
],
[
"Wang",
"Jinyin",
""
]
] |
2404.17334 | Luis Anchordoqui | Luis A. Anchordoqui, Ignatios Antoniadis, Dieter Lust, Neena T. Noble,
Jorge F. Soriano | From infinite to infinitesimal: Using the Universe as a dataset to probe
Casimir corrections to the vacuum energy from fields inhabiting the dark
dimension | Appendix added; conclusions unchanged | Phys. Dark Univ. 46 (2024) 101715 | 10.1016/j.dark.2024.101715 | MPP-2024-91; LMU-ASC 05/24 | astro-ph.CO hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Promptly after high-resolution experiments harbinger the field of precision
cosmology low- and high-redshift observations abruptly gave rise to a tension
in the measurement of the present-day expansion rate of the Universe ($H_0$)
and the clustering of matter ($S_8$). The statistically significant
discrepancies between the locally measured values of $H_0$ and $S_8$ and the
ones inferred from observations of the cosmic microwave background assuming the
canonical $\Lambda$ cold dark matter (CDM) cosmological model have become a new
cornerstone of theoretical physics. $\Lambda_s$CDM is one of the many beyond
Standard Model setups that have been proposed to simultaneously resolve the
cosmological tensions. This setup relies on an empirical conjecture, which
postulates that $\Lambda$ switched sign (from negative to positive) at a
critical redshift $z_c \sim 2$. We reexamine a stringy model that can describe
the transition in the vacuum energy hypothesized in $\Lambda_s$CDM. The model
makes use of the Casimir forces driven by fields inhabiting the incredible bulk
of the dark dimension scenario. Unlike the $\Lambda_s$CDM setup the model
deviates from $\Lambda$CDM in the early universe due to the existence of
relativistic neutrino-like species. Using the Boltzmann solver CLASS in
combination with MontePython we confront predictions of the stringy model to
experimental data (from the Planck mission, Pantheon+ supernova type Ia, BAO,
and KiDS-1000). We show that the string-inspired model provides a satisfactory
fit to the data and can resolve the cosmological tensions.
| [
{
"version": "v1",
"created": "Fri, 26 Apr 2024 11:32:27 GMT"
},
{
"version": "v2",
"created": "Wed, 19 Jun 2024 12:30:10 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Oct 2024 08:43:29 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Anchordoqui",
"Luis A.",
""
],
[
"Antoniadis",
"Ignatios",
""
],
[
"Lust",
"Dieter",
""
],
[
"Noble",
"Neena T.",
""
],
[
"Soriano",
"Jorge F.",
""
]
] |
2404.18110 | Shangkun Weng | Shangkun Weng and Zhouping Xin | Some three dimensional smooth transonic flows for the steady Euler
equations with an external force | 59 pages | null | null | null | math.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We establish the existence and uniqueness of some smooth accelerating
transonic flows governed by the three dimensional steady compressible Euler
equations with an external force in cylinders with arbitrary cross sections,
which include both irrotational flows and Beltrami flows with nonuniform
proportionality factors. One of the key ingredients in the analysis of smooth
transonic irrotational flows is the well-posedness theory of classical
solutions in $H^4$ to a linear elliptic-hyperbolic mixed second order
differential equation of Keldysh type in cylinders with mixed boundary
conditions. This is achieved by extending the problem to an auxiliary linear
elliptic-hyberbolic-elliptic mixed problem in a longer cylinder where the
governing equation becomes elliptic at the exit of the new cylinder, so that
one can use the multiplier method and the cut-off techniques to derive the
$H^2$ and higher order estimates in transonic regions. It is further shown that
the energy estimate can be closed in the $H^4$ framework. For smooth transonic
Beltrami flows, we solve a transport equation for the proportionality factor
and a type-changing enlarged deformation-curl system with mixed boundary
conditions. The compatibility conditions for the $H^4$ estimate to the enlarged
deformation-curl system near the intersection between the entrance and the
cylinder wall play a crucial role in the analysis.
| [
{
"version": "v1",
"created": "Sun, 28 Apr 2024 08:13:06 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 02:44:30 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Weng",
"Shangkun",
""
],
[
"Xin",
"Zhouping",
""
]
] |
2405.00533 | David Wagner | David Wagner, Masoud Shokri, Dirk H. Rischke | Damping of spin waves | 13 pages, 1 figure | Phys.Rev.Res. 6 (2024) 4, 043103 | 10.1103/PhysRevResearch.6.043103 | null | nucl-th hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We show that, in ideal-spin hydrodynamics, the components of the spin tensor
follow damped wave equations. The damping rate is related to nonlocal
collisions of the particles in the fluid, which enter at first order in $\hbar$
in a semi-classical expansion. This rate provides an estimate for the timescale
of spin equilibration and is computed by considering a system of spin-1/2
fermions interacting via a quartic self-interaction as well as via (screened)
one-gluon exchange. It is found that the relaxation times of the components of
the spin tensor can become very large compared to the usual dissipative
timescales of the system. Our results suggest that the spin degrees of freedom
in a heavy-ion collision may not be in equilibrium by the time of freeze-out,
and thus should be treated dynamically.
| [
{
"version": "v1",
"created": "Wed, 1 May 2024 14:06:11 GMT"
},
{
"version": "v2",
"created": "Wed, 6 Nov 2024 20:46:10 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Wagner",
"David",
""
],
[
"Shokri",
"Masoud",
""
],
[
"Rischke",
"Dirk H.",
""
]
] |
2405.00695 | Giulia D'Addato | Giulia d'Addato, Ruggero Carli, Eurico Pedrosa, Artur Pereira, Luigi
Palopoli, Daniele Fontanelli | Joint torques prediction of a robotic arm using neural networks | 6 pages, 5 figures, submitted to CASE 2024 | null | null | null | cs.RO cs.LG cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurate dynamic models are crucial for many robotic applications.
Traditional approaches to deriving these models are based on the application of
Lagrangian or Newtonian mechanics. Although these methods provide a good
insight into the physical behaviour of the system, they rely on the exact
knowledge of parameters such as inertia, friction and joint flexibility. In
addition, the system is often affected by uncertain and nonlinear effects, such
as saturation and dead zones, which can be difficult to model. A popular
alternative is the application of Machine Learning (ML) techniques - e.g.,
Neural Networks (NNs) - in the context of a "black-box" methodology. This paper
reports on our experience with this approach for a real-life 6 degrees of
freedom (DoF) manipulator. Specifically, we considered several NN
architectures: single NN, multiple NNs, and cascade NN. We compared the
performance of the system by using different policies for selecting the NN
hyperparameters. Our experiments reveal that the best accuracy and performance
are obtained by a cascade NN, in which we encode our prior physical knowledge
about the dependencies between joints, complemented by an appropriate
optimisation of the hyperparameters.
| [
{
"version": "v1",
"created": "Thu, 28 Mar 2024 09:38:26 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"d'Addato",
"Giulia",
""
],
[
"Carli",
"Ruggero",
""
],
[
"Pedrosa",
"Eurico",
""
],
[
"Pereira",
"Artur",
""
],
[
"Palopoli",
"Luigi",
""
],
[
"Fontanelli",
"Daniele",
""
]
] |
2405.01076 | Robert Hahn | Robert Hahn, Erik Schnaubelt, Mariusz Wozniak, Christophe Geuzaine,
Sebastian Sch\"ops | Mortar Thin Shell Approximation for Analysis of Superconducting
Accelerator Magnets | Pre-submission version (preprint). Presented at and submitted to the
proceedings of "The 15th International Conference on Scientific Computing in
Electrical Engineering" (SCEE 2024), March 4-8, 2024, Darmstadt, Germany | null | null | null | cs.CE physics.comp-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Thin layers can lead to unfavorable meshes in a finite element (FE) analysis.
Thin shell approximations (TSAs) avoid this issue by removing the need for a
mesh of the thin layer while approximating the physics across the layer by an
interface condition. Typically, a TSA requires the mesh of both sides of the
TSA interface to be conforming. To alleviate this requirement, we propose to
combine mortar methods and TSAs for solving the heat equation. The mortar TSA
method's formulation is derived and enables an independent discretization of
the subdomains on the two sides of the TSA depending on their accuracy
requirements. The method is verified by comparison with a reference FE solution
of a thermal model problem of a simplified superconducting accelerator magnet.
| [
{
"version": "v1",
"created": "Thu, 2 May 2024 08:13:42 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 12:17:26 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Hahn",
"Robert",
""
],
[
"Schnaubelt",
"Erik",
""
],
[
"Wozniak",
"Mariusz",
""
],
[
"Geuzaine",
"Christophe",
""
],
[
"Schöps",
"Sebastian",
""
]
] |
2405.01090 | Masatoshi Tateno | Masatoshi Tateno, Takuma Yagi, Ryosuke Furuta, Yoichi Sato | Learning Multiple Object States from Actions via Large Language Models | This is an accepted WACV 2025 paper. Project page:
https://masatate.github.io/ObjStatefromAction.github.io/ | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recognizing the states of objects in a video is crucial in understanding the
scene beyond actions and objects. For instance, an egg can be raw, cracked, and
whisked while cooking an omelet, and these states can coexist simultaneously
(an egg can be both raw and whisked). However, most existing research assumes a
single object state change (e.g., uncracked -> cracked), overlooking the
coexisting nature of multiple object states and the influence of past states on
the current state. We formulate object state recognition as a multi-label
classification task that explicitly handles multiple states. We then propose to
learn multiple object states from narrated videos by leveraging large language
models (LLMs) to generate pseudo-labels from the transcribed narrations,
capturing the influence of past states. The challenge is that narrations mostly
describe human actions in the video but rarely explain object states.
Therefore, we use the LLMs knowledge of the relationship between actions and
states to derive the missing object states. We further accumulate the derived
object states to consider past state contexts to infer current object state
pseudo-labels. We newly collect a dataset called the Multiple Object States
Transition (MOST) dataset, which includes manual multi-label annotation for
evaluation purposes, covering 60 object states across six object categories.
Experimental results show that our model trained on LLM-generated pseudo-labels
significantly outperforms strong vision-language models, demonstrating the
effectiveness of our pseudo-labeling framework that considers past context via
LLMs.
| [
{
"version": "v1",
"created": "Thu, 2 May 2024 08:43:16 GMT"
},
{
"version": "v2",
"created": "Wed, 6 Nov 2024 11:44:50 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Tateno",
"Masatoshi",
""
],
[
"Yagi",
"Takuma",
""
],
[
"Furuta",
"Ryosuke",
""
],
[
"Sato",
"Yoichi",
""
]
] |
2405.01331 | Mikhail Khenner | Mikhail Khenner | On Nanowire Morphological Instability and Pinch-Off by Surface
Electromigration | null | null | null | null | cond-mat.mtrl-sci cond-mat.mes-hall math-ph math.MP physics.app-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Surface diffusion and surface electromigration may lead to a morphological
instability of thin solid films and nanowires. In this paper two nonlinear
analyses of a morphological instability are developed for a single-crystal
cylindrical nanowire that is subjected to an axial current. These treatments
extend the conventional linear stability analyses without surface
electromigration, that manifest a Rayleigh-Plateau instability. A weakly
nonlinear analysis is done slightly above the Rayleigh-Plateau (longwave)
instability threshold. It results in a one-dimensional Sivashinsky amplitude
equation that describes a blow-up of a surface perturbation amplitude in a
finite time. This is a signature of a pinching singularity of a cylinder
radius, which leads to a wire separation into a disjoint segments. The time-
and electric field-dependent dimensions of the focusing self-similar amplitude
profile approaching a blow-up are characterized via the scaling analysis. Also,
a weakly nonlinear multi-scale analysis is done at the arbitrary distance above
a longwave or a shortwave instability threshold. The time- and electric
field-dependent Fourier amplitudes of the major instability modes are derived
and characterized.
| [
{
"version": "v1",
"created": "Thu, 2 May 2024 14:35:47 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 16:41:52 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Khenner",
"Mikhail",
""
]
] |
2405.01431 | Francesco Anna Mele | Francesco Anna Mele, Antonio Anna Mele, Lennart Bittel, Jens Eisert,
Vittorio Giovannetti, Ludovico Lami, Lorenzo Leone, Salvatore F.E. Oliviero | Learning quantum states of continuous variable systems | In the fourth version of the paper, we corrected a typo in Lemma S55 | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Quantum state tomography, aimed at deriving a classical description of an
unknown state from measurement data, is a fundamental task in quantum physics.
In this work, we analyse the ultimate achievable performance of tomography of
continuous-variable systems, such as bosonic and quantum optical systems. We
prove that tomography of these systems is extremely inefficient in terms of
time resources, much more so than tomography of finite-dimensional systems: not
only does the minimum number of state copies needed for tomography scale
exponentially with the number of modes, but it also exhibits a dramatic scaling
with the trace-distance error, even for low-energy states, in stark contrast
with the finite-dimensional case. On a more positive note, we prove that
tomography of Gaussian states is efficient. To accomplish this, we answer a
fundamental question for the field of continuous-variable quantum information:
if we know with a certain error the first and second moments of an unknown
Gaussian state, what is the resulting trace-distance error that we make on the
state? Lastly, we demonstrate that tomography of non-Gaussian states prepared
through Gaussian unitaries and a few local non-Gaussian evolutions is efficient
and experimentally feasible.
| [
{
"version": "v1",
"created": "Thu, 2 May 2024 16:19:55 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Sep 2024 19:48:12 GMT"
},
{
"version": "v3",
"created": "Mon, 7 Oct 2024 10:46:02 GMT"
},
{
"version": "v4",
"created": "Thu, 7 Nov 2024 09:33:36 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Mele",
"Francesco Anna",
""
],
[
"Mele",
"Antonio Anna",
""
],
[
"Bittel",
"Lennart",
""
],
[
"Eisert",
"Jens",
""
],
[
"Giovannetti",
"Vittorio",
""
],
[
"Lami",
"Ludovico",
""
],
[
"Leone",
"Lorenzo",
""
],
[
"Oliviero",
"Salvatore F. E.",
""
]
] |
2405.01498 | Kaushik Chatterjee Dr. | Kaushik Chatterjee, Santanu Mondal, Chandra B. Singh, and Mutsumi
Sugizaki | {\it Insight}-HXMT View of the BHC Swift J1727.8-1613 during its
outburst in 2023 | Accepted for Publication in ApJ (21 pages, 9 figures, 5 tables) | null | 10.3847/1538-4357/ad8dc4 | null | astro-ph.HE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The transient Galactic black hole candidate Swift\,J1727.8-1613 went through
an outburst for the very first time in August 2023 and lasted for almost 6
months. We study the timing and spectral properties of this source using
publicly available archival {\it Insight}-HXMT data for the first 10
observation IDs that last from MJD 60181 to 60198 with a total of 92 exposures
for all three energy bands. We have detected quasi-periodic oscillations (QPOs)
in a frequency range of $0.21 \pm 0.01$ - $1.86 \pm 0.01$ Hz by fitting the
power density spectrum. Based on the model-fitted parameters and properties of
the QPOs, we classify them as type-C in nature. We also conclude that the
origin of the QPOs could be the shock instabilities in the transonic advective
accretion flows around black holes. The spectral analysis was performed using
simultaneous data from the three on-board instruments LE, ME, and HE of
\textit{Insight}-HXMT in the broad energy band of $2-150 $ keV. To achieve the
best fit, spectral fitting required a combination of models e.g. interstellar
absorption, power-law, multi-color disk-blackbody continuum, Gaussian
emission/absorption, and reflection by neutral material. From the spectral
properties, we found that the source was in an intermediate state at the start
of the analysis period and was transitioning to the softer states. The inner
edge of the accretion disk moved inward in progressive days following the
spectral nature. We found that the source has a high inclination of
$78^\circ-86^\circ$. The hydrogen column density from the model fitting varied
in the range of $(0.12 \pm 0.02 - 0.39 \pm 0.08)\times10^{22}$ cm$^{-2}$.
| [
{
"version": "v1",
"created": "Thu, 2 May 2024 17:28:01 GMT"
},
{
"version": "v2",
"created": "Sun, 12 May 2024 05:29:38 GMT"
},
{
"version": "v3",
"created": "Thu, 7 Nov 2024 03:47:21 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Chatterjee",
"Kaushik",
""
],
[
"Mondal",
"Santanu",
""
],
[
"Singh",
"Chandra B.",
""
],
[
"Sugizaki",
"Mutsumi",
""
]
] |
2405.02395 | Alessandro Lella | Alessandro Lella, Francesca Calore, Pierluca Carenza, Christopher
Eckner, Maurizio Giannotti, Giuseppe Lucente, Alessandro Mirizzi | Probing protoneutron stars with gamma-ray axionscopes | 25 pages, 11 figures, 3 tables. Matches the version published on
JCAP. Changes: we added further discussion of possible perspectives beyond
the Fermi-LAT operational period | JCAP 11 (2024) 009 | 10.1088/1475-7516/2024/11/009 | LAPTH-024/24, BARI-TH/769-24 | hep-ph astro-ph.HE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Axion-like particles (ALPs) coupled to nucleons can be efficiently produced
in the interior of protoneutron stars (PNS) during supernova (SN) explosions.
If these ALPs are also coupled to photons they can convert into gamma rays in
the Galactic magnetic field. This SN-induced gamma-ray burst can be observable
by gamma-ray telescopes like ${\textit Fermi}$-LAT if the SN is in the field of
view of the detector. We show that the observable gamma-ray spectrum is
sensitive to the production processes in the SN core. In particular, if the
nucleon-nucleon bremsstrahlung is the dominant axion production channel, one
expects a thermal spectrum with average energy $E_a \simeq 50$ MeV. In this
case the gamma-ray spectrum observation allows for the reconstruction of the
PNS temperature. In case of a sizable pion abundance in the SN core, one
expects a second spectral component peaked at $E_a\simeq 200$ MeV due to axion
pionic processes. We demonstrate that, through a dedicated LAT analysis, we can
detect the presence of this pionic contribution, showing that the detection of
the spectral shape of the gamma-ray signal represents a unique probe of the
pion abundance in the PNS.
| [
{
"version": "v1",
"created": "Fri, 3 May 2024 18:00:01 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 16:10:37 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Lella",
"Alessandro",
""
],
[
"Calore",
"Francesca",
""
],
[
"Carenza",
"Pierluca",
""
],
[
"Eckner",
"Christopher",
""
],
[
"Giannotti",
"Maurizio",
""
],
[
"Lucente",
"Giuseppe",
""
],
[
"Mirizzi",
"Alessandro",
""
]
] |
2405.03497 | Matthias Fuchs | Niklas Grimm, Martin von Bischopinck, Andreas Zumbusch, and Matthias
Fuchs | Long ranged stress correlations in the hard sphere liquid | 15 pages, 18 figures | J. Chem. Phys. 161, 144118 (2024) | 10.1063/5.0225890 | null | cond-mat.soft cond-mat.stat-mech | http://creativecommons.org/licenses/by/4.0/ | The smooth emergence of shear elasticity is an hallmark of the liquid to
glass transition. In a liquid, viscous stresses arise from local structural
rearrangements. In the solid, Eshelby has shown that stresses around an
inclusion decay as a power law $r^{-D}$, where $D$ is the dimension of the
system. We study glass-forming hard sphere fluids by simulation and observe the
emergence of the unscreened power-law Eshelby pattern in the stress
correlations of the isotropic liquid state. By a detailed tensorial analysis,
we show that the fluctuating force field, viz.~the divergence of the stress
field, relaxes to zero with time in all states, while the shear stress
correlations develop spatial power-law structures inside regions that grow with
longitudinal and transverse sound speeds; we observe the predicted exponents
$r^{-D}$ and $r^{-D-2}$. In Brownian systems, shear stresses relax diffusively
within these regions, with the diffusion coefficient determined by the shear
modulus and the friction coefficient.
| [
{
"version": "v1",
"created": "Mon, 6 May 2024 14:07:05 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Grimm",
"Niklas",
""
],
[
"von Bischopinck",
"Martin",
""
],
[
"Zumbusch",
"Andreas",
""
],
[
"Fuchs",
"Matthias",
""
]
] |
2405.03655 | Mohammad Mortezaei Nobahari | Mohammad Mortezaei Nobahari, Carmine Autieri | Spin-Hall conductivity and optical characteristics of noncentrosymmetric
quantum spin Hall insulators: the case of PbBiI | 7 pages and 9 figures | Scientific Reports volume 14, Article number: 27041 (2024) | 10.1038/s41598-024-77403-9 | null | cond-mat.mes-hall | http://creativecommons.org/licenses/by/4.0/ | Quantum spin Hall insulators have attracted significant attention in recent
years. Understanding the optical properties and spin Hall effect in these
materials is crucial for technological advancements. In this study, we present
theoretical analyses to explore the optical properties, Berry curvature and
spin Hall conductivity of pristine and perturbed PbBiI using the linear
combination of atomic orbitals and the Kubo formula. The system is not
centrosymmetric and it is hosting at the same time Rashba spin-splitting and
quantized spin Hall conductivity. Our calculations reveal that the electronic
structure can be modified using staggered exchange fields and electric fields,
leading to changes in the optical properties. Additionally, the spin Berry
curvature and spin Hall conductivity are investigated as a function of the
energy and temperature. The results indicate that due to the small dynamical
spin Hall conductivity, generating an ac spin current in PbBiI requires the use
of external magnetic fields or magnetic materials.
| [
{
"version": "v1",
"created": "Mon, 6 May 2024 17:28:31 GMT"
},
{
"version": "v2",
"created": "Thu, 22 Aug 2024 08:45:48 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Nobahari",
"Mohammad Mortezaei",
""
],
[
"Autieri",
"Carmine",
""
]
] |
2405.03987 | Guanghao Wei | Guanghao Wei and Yining Huang and Chenru Duan and Yue Song and Yuanqi
Du | Navigating Chemical Space with Latent Flows | Accepted for presentation at NeurIPS 2024 | null | null | null | cs.LG physics.chem-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent progress of deep generative models in the vision and language domain
has stimulated significant interest in more structured data generation such as
molecules. However, beyond generating new random molecules, efficient
exploration and a comprehensive understanding of the vast chemical space are of
great importance to molecular science and applications in drug design and
materials discovery. In this paper, we propose a new framework, ChemFlow, to
traverse chemical space through navigating the latent space learned by molecule
generative models through flows. We introduce a dynamical system perspective
that formulates the problem as learning a vector field that transports the mass
of the molecular distribution to the region with desired molecular properties
or structure diversity. Under this framework, we unify previous approaches on
molecule latent space traversal and optimization and propose alternative
competing methods incorporating different physical priors. We validate the
efficacy of ChemFlow on molecule manipulation and single- and multi-objective
molecule optimization tasks under both supervised and unsupervised molecular
discovery settings. Codes and demos are publicly available on GitHub at
https://github.com/garywei944/ChemFlow.
| [
{
"version": "v1",
"created": "Tue, 7 May 2024 03:55:57 GMT"
},
{
"version": "v2",
"created": "Wed, 8 May 2024 01:34:25 GMT"
},
{
"version": "v3",
"created": "Wed, 6 Nov 2024 22:13:26 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Wei",
"Guanghao",
""
],
[
"Huang",
"Yining",
""
],
[
"Duan",
"Chenru",
""
],
[
"Song",
"Yue",
""
],
[
"Du",
"Yuanqi",
""
]
] |
2405.04304 | Jonathan Mamou | Jonathan Mamou and Oren Pereg and Daniel Korat and Moshe Berchansky
and Nadav Timor and Moshe Wasserblat and Roy Schwartz | Dynamic Speculation Lookahead Accelerates Speculative Decoding of Large
Language Models | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Speculative decoding is commonly used for reducing the inference latency of
large language models. Its effectiveness depends highly on the speculation
lookahead (SL)-the number of tokens generated by the draft model at each
iteration. In this work we show that the common practice of using the same SL
for all iterations (static SL) is suboptimal. We introduce DISCO (DynamIc
SpeCulation lookahead Optimization), a novel method for dynamically selecting
the SL. Our experiments with four datasets show that DISCO reaches an average
speedup of 10% compared to the best static SL baseline, while generating the
exact same text.
| [
{
"version": "v1",
"created": "Tue, 7 May 2024 13:27:52 GMT"
},
{
"version": "v2",
"created": "Tue, 18 Jun 2024 12:34:35 GMT"
},
{
"version": "v3",
"created": "Wed, 19 Jun 2024 08:54:51 GMT"
},
{
"version": "v4",
"created": "Sun, 23 Jun 2024 13:46:23 GMT"
},
{
"version": "v5",
"created": "Thu, 7 Nov 2024 12:59:39 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Mamou",
"Jonathan",
""
],
[
"Pereg",
"Oren",
""
],
[
"Korat",
"Daniel",
""
],
[
"Berchansky",
"Moshe",
""
],
[
"Timor",
"Nadav",
""
],
[
"Wasserblat",
"Moshe",
""
],
[
"Schwartz",
"Roy",
""
]
] |
2405.04906 | Navid Zehtabiyan-Rezaie | Ali Amarloo, Navid Zehtabiyan-Rezaie, Mahdi Abkar | A progressive data-augmented RANS model for enhanced wind-farm
simulations | A. Amarloo and N. Zehtabiyan-Rezaie contributed equally to this study | null | null | null | physics.flu-dyn | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The development of advanced simulation tools is essential, both presently and
in the future, for improving wind-energy design strategies, paving the way for
a complete transition to sustainable solutions. The Reynolds-averaged
Navier-Stokes (RANS) models are pivotal in enhancing our comprehension of the
complex flow within and around wind farms and, hence, improving their capacity
to accurately model turbulence within this context is a vital research goal.
The enhancement is essential for a precise prediction of wake recovery and for
capturing intricate flow phenomena such as secondary flows of Prandtl's second
kind behind the turbines. To reach these objectives, here, we propose a
progressive data-augmentation approach. We first incorporate the
turbine-induced forces in the turbulent kinetic energy equation of the widely
used $k-\omega\text{SST}$ model. Afterward, we utilize data from large-eddy
simulations to progressively enhance the Reynolds-stress prediction of this
baseline model, accurately capturing the evolution of eddy viscosity in the
wake, as well as the emergence of secondary flows. We then apply the optimized
model to two unseen cases with distinct layouts and conduct a comparative
analysis focusing on the obtained quantities such as normalized streamwise
velocity deficit, turbulence intensity, and power output. We also examine the
success rate of the augmented model in predicting the secondary flows in the
wake region. We also evaluate the performance of the augmented model in
predicting wake characteristics by comparing it with wind-tunnel measurement
data. Our comparisons and validations demonstrate the superior performance of
the progressive data-augmented model over the standard version in all cases
considered in this study.
| [
{
"version": "v1",
"created": "Wed, 8 May 2024 09:24:10 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 07:28:31 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Amarloo",
"Ali",
""
],
[
"Zehtabiyan-Rezaie",
"Navid",
""
],
[
"Abkar",
"Mahdi",
""
]
] |
2405.05038 | Sreejith Nair | Sreejith Nair, Aditya Vijaykumar, Sudipta Sarkar | Bounds on the charge of the graviton using gravitational wave
observations | 9 pages, 3 figures, Accepted for publication in JCAP | null | null | LIGO DCC P2400160 | gr-qc astro-ph.HE hep-ph hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | If the graviton possesses a non-zero charge $q_g$, gravitational waves (GW)
originating from astrophysical sources would experience an additional time
delay due to intergalactic magnetic fields. This would result in a modification
of the phase evolution of the observed GW signal similar to the effect induced
by a massive graviton. As a result, we can reinterpret the most recent upper
limits on the graviton's mass as constraints on the joint mass-charge parameter
space, finding $|q_g|/{e} < 3\times 10^{-34}$ where $e$ represents the charge
of an electron. Additionally, we illustrate that a charged graviton would
introduce a constant phase difference in the gravitational waves detected by
two spatially separated GW detectors due to the Aharonov-Bohm effect. Using the
non-observation of such a phase difference for the GW event GW190814, we
establish a mass-independent constraint $|q_g|/e < 2\times 10^{-26}$. To the
best of our knowledge, our results constitute the first-ever bounds on the
charge of the graviton. We also discuss various caveats involved in our
measurements and prospects for strengthening these bounds with future GW
observations.
| [
{
"version": "v1",
"created": "Wed, 8 May 2024 13:11:02 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 05:07:17 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Nair",
"Sreejith",
""
],
[
"Vijaykumar",
"Aditya",
""
],
[
"Sarkar",
"Sudipta",
""
]
] |
2405.06481 | Evan Yerger | Evan L. Yerger, Matthew W. Kunz, Archie F. A. Bott, Anatoly Spitkovsky | Collisionless conduction in a high-beta plasma: a collision operator for
whistler turbulence | null | null | null | null | astro-ph.HE physics.plasm-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The regulation of electron heat transport in high-beta, weakly collisional,
magnetized plasma is investigated. A temperature gradient oriented along a mean
magnetic field can induce a kinetic heat-flux-driven whistler instability
(HWI), which back-reacts on the transport by scattering electrons and impeding
their flow. Previous analytical and numerical studies have shown that the heat
flux for the saturated HWI scales as the inverse of electron beta. These
numerical studies, however, had limited scale separation and consequently large
fluctuation amplitudes, which calls into question their relevance at
astrophysical scales. To this end, we perform a series of particle-in-cell
simulations of the HWI across a range of electron beta and temperature-gradient
length scales under two different physical setups. The saturated heat flux in
all of our simulations follows the expected inverse-electron-beta scaling,
supporting the robustness of the result. We also use our simulation results to
develop and implement several methods to construct an effective collision
operator for whistler turbulence. The results point to an issue with the
standard quasi-linear explanation of HWI saturation, which is analogous to the
well-known 90-degree scattering problem in the cosmic ray community. Despite
this limitation, the methods developed here can serve as a blueprint for future
work seeking to characterize the effective collisionality caused by kinetic
instabilities.
| [
{
"version": "v1",
"created": "Fri, 10 May 2024 13:58:05 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 16:17:39 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Yerger",
"Evan L.",
""
],
[
"Kunz",
"Matthew W.",
""
],
[
"Bott",
"Archie F. A.",
""
],
[
"Spitkovsky",
"Anatoly",
""
]
] |
2405.06531 | Antonio Picozzi | Lucas Zanaglia, Josselin Garnier, Sergio Rica, Robin Kaiser, Stefano
Wabnitz, Claire Michel, Valerie Doya, Antonio Picozzi | Bridging Rayleigh-Jeans and Bose-Einstein condensation of a guided fluid
of light with positive and negative temperatures | 21 pages, 12 figures | null | null | null | physics.optics nlin.PS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the free propagation geometry of a light beam (or fluid of light)
in a multimode waveguide. As a result of the effective photon-photon
interactions, the photon fluid thermalizes to an equilibrium state during its
conservative propagation. In this configuration, Rayleigh-Jeans (RJ)
thermalization and condensation of classical light waves have been recently
observed experimentally in graded index multimode optical fibers characterized
by a 2D parabolic trapping potential. As well-known, the properties of RJ
condensation differ substantially from those of Bose-Einstein (BE)
condensation: The condensate fraction decreases quadratically with the
temperature for BE condensation, while it decreases linearly for RJ
condensation. Furthermore, for quantum particles the heat capacity tends to
zero at small temperatures, and it takes a constant value in the classical
particle limit at high temperatures. This is in contrast with classical RJ
waves, where the specific heat takes a constant value at small temperatures,
and tends to vanish above the condensation transition in the normal
(uncondensed) state. Here, we reconcile the thermodynamic properties of BE and
RJ condensation: By introducing a frequency cut-off inherent to light
propagation in a waveguide, we derive generalized expressions of the
thermodynamic properties that include the RJ and BE limits as particular cases.
We extend the approach to encompass negative temperatures. In contrast to
positive temperatures, the specific heat does not display a singular behavior
at negative temperatures, reflecting the non-critical nature of the transition
to a macroscopic population of the highest energy level. Our work contributes
to understanding the quantum-to-classical crossover in the equilibrium
properties of light, within a versatile experimental platform based on
nonlinear optical propagation in multimode waveguides.
| [
{
"version": "v1",
"created": "Fri, 10 May 2024 15:20:07 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 18:55:27 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Zanaglia",
"Lucas",
""
],
[
"Garnier",
"Josselin",
""
],
[
"Rica",
"Sergio",
""
],
[
"Kaiser",
"Robin",
""
],
[
"Wabnitz",
"Stefano",
""
],
[
"Michel",
"Claire",
""
],
[
"Doya",
"Valerie",
""
],
[
"Picozzi",
"Antonio",
""
]
] |
2405.08107 | Juri Poutanen | Juri Poutanen, Sergey S. Tsygankov, Victor Doroshenko, Sofia V.
Forsblom, Peter Jenke, Philip Kaaret, Andrei V. Berdyugin, Dmitry Blinov,
Vadim Kravtsov, Ioannis Liodakis, Anastasia Tzouvanou, Alessandro Di Marco,
Jeremy Heyl, Fabio La Monaca, Alexander A. Mushtukov, George G. Pavlov,
Alexander Salganik, Alexandra Veledina, Martin C. Weisskopf, Silvia Zane,
Vladislav Loktev, Valery F. Suleimanov, Colleen Wilson-Hodge, Svetlana V.
Berdyugina, Masato Kagitani, Vilppu Piirola, Takeshi Sakanoi, Ivan Agudo,
Lucio A. Antonelli, Matteo Bachetti, Luca Baldini, Wayne H. Baumgartner,
Ronaldo Bellazzini, Stefano Bianchi, Stephen D. Bongiorno, Raffaella Bonino,
Alessandro Brez, Niccolo Bucciantini, Fiamma Capitanio, Simone Castellano,
Elisabetta Cavazzuti, Chien-Ting Chen, Stefano Ciprini, Enrico Costa,
Alessandra De Rosa, Ettore Del Monte, Laura Di Gesu, Niccolo Di Lalla,
Immacolata Donnarumma, Michal Dovciak, Steven R. Ehlert, Teruaki Enoto, Yuri
Evangelista, Sergio Fabiani, Riccardo Ferrazzoli, Javier A. Garcia, Shuichi
Gunji, Kiyoshi Hayashida, Wataru Iwakiri, Svetlana G. Jorstad, Vladimir
Karas, Fabian Kislat, Takao Kitaguchi, Jeffery J. Kolodziejczak, Luca
Latronico, Simone Maldera, Alberto Manfreda, Frederic Marin, Andrea
Marinucci, Alan P. Marscher, Herman L. Marshall, Francesco Massaro, Giorgio
Matt, Ikuyuki Mitsuishi, Tsunefumi Mizuno, Fabio Muleri, Michela Negro,
Chi-Yung Ng, Stephen L. O'Dell, Nicola Omodei, Chiara Oppedisano, Alessandro
Papitto, Abel L. Peirson, Matteo Perri, Melissa Pesce-Rollins, Pierre-Olivier
Petrucci, Maura Pilia, Andrea Possenti, Simonetta Puccetti, Brian D. Ramsey,
John Rankin, Ajay Ratheesh, Oliver J. Roberts, Roger W. Romani, Carmelo Sgro,
Patrick Slane, Paolo Soffitta, Gloria Spandre, Douglas A. Swartz, Toru
Tamagawa, Fabrizio Tavecchio, Roberto Taverna, Yuzuru Tawara, Allyn F.
Tennant, Nicholas E. Thomas, Francesco Tombesi, Alessio Trois, Roberto
Turolla, Jacco Vink, Kinwah Wu, Fei Xie | Studying geometry of the ultraluminous X-ray pulsar Swift J0243.6+6124
using X-ray and optical polarimetry | 13 pages, 10 figures, version published in A&A | A&A 691, A123 (2024) | 10.1051/0004-6361/202450696 | null | astro-ph.HE astro-ph.SR | http://creativecommons.org/licenses/by/4.0/ | Discovery of pulsations from a number of ULXs proved that accretion onto
neutron stars can produce luminosities exceeding the Eddington limit by several
orders of magnitude. The conditions necessary to achieve such high luminosities
as well as the exact geometry of the accretion flow in the neutron star
vicinity are, however, a matter of debate. The pulse phase-resolved
polarization measurements that became possible with the launch of the Imaging
X-ray Polarimetry Explorer (IXPE) can be used to determine the pulsar geometry
and its orientation relative to the orbital plane. They provide an avenue to
test different theoretical models of ULX pulsars. In this paper we present the
results of three IXPE observations of the first Galactic ULX pulsar Swift
J0243.6+6124 during its 2023 outburst. We find strong variations in the
polarization characteristics with the pulsar phase. The average polarization
degree increases from about 5% to 15% as the flux dropped by a factor of three
in the course of the outburst. The polarization angle (PA) as a function of the
pulsar phase shows two peaks in the first two observations, but changes to a
characteristic sawtooth pattern in the remaining data set. This is not
consistent with a simple rotating vector model. Assuming the existence of an
additional constant polarized component, we were able to fit the three
observations with a common rotating vector model and obtain constraints on the
pulsar geometry. In particular, we find the pulsar angular momentum inclination
with respect to the line of sight of 15-40 deg, the magnetic obliquity of 60-80
deg, and the pulsar spin position angle of -50 deg, which significantly differs
from the constant component PA of about 10 deg. Combining these X-ray
measurements with the optical PA, we find evidence for at least a 30 deg
misalignment between the pulsar angular momentum and the binary orbital axis.
| [
{
"version": "v1",
"created": "Mon, 13 May 2024 18:40:50 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 09:33:17 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Poutanen",
"Juri",
""
],
[
"Tsygankov",
"Sergey S.",
""
],
[
"Doroshenko",
"Victor",
""
],
[
"Forsblom",
"Sofia V.",
""
],
[
"Jenke",
"Peter",
""
],
[
"Kaaret",
"Philip",
""
],
[
"Berdyugin",
"Andrei V.",
""
],
[
"Blinov",
"Dmitry",
""
],
[
"Kravtsov",
"Vadim",
""
],
[
"Liodakis",
"Ioannis",
""
],
[
"Tzouvanou",
"Anastasia",
""
],
[
"Di Marco",
"Alessandro",
""
],
[
"Heyl",
"Jeremy",
""
],
[
"La Monaca",
"Fabio",
""
],
[
"Mushtukov",
"Alexander A.",
""
],
[
"Pavlov",
"George G.",
""
],
[
"Salganik",
"Alexander",
""
],
[
"Veledina",
"Alexandra",
""
],
[
"Weisskopf",
"Martin C.",
""
],
[
"Zane",
"Silvia",
""
],
[
"Loktev",
"Vladislav",
""
],
[
"Suleimanov",
"Valery F.",
""
],
[
"Wilson-Hodge",
"Colleen",
""
],
[
"Berdyugina",
"Svetlana V.",
""
],
[
"Kagitani",
"Masato",
""
],
[
"Piirola",
"Vilppu",
""
],
[
"Sakanoi",
"Takeshi",
""
],
[
"Agudo",
"Ivan",
""
],
[
"Antonelli",
"Lucio A.",
""
],
[
"Bachetti",
"Matteo",
""
],
[
"Baldini",
"Luca",
""
],
[
"Baumgartner",
"Wayne H.",
""
],
[
"Bellazzini",
"Ronaldo",
""
],
[
"Bianchi",
"Stefano",
""
],
[
"Bongiorno",
"Stephen D.",
""
],
[
"Bonino",
"Raffaella",
""
],
[
"Brez",
"Alessandro",
""
],
[
"Bucciantini",
"Niccolo",
""
],
[
"Capitanio",
"Fiamma",
""
],
[
"Castellano",
"Simone",
""
],
[
"Cavazzuti",
"Elisabetta",
""
],
[
"Chen",
"Chien-Ting",
""
],
[
"Ciprini",
"Stefano",
""
],
[
"Costa",
"Enrico",
""
],
[
"De Rosa",
"Alessandra",
""
],
[
"Del Monte",
"Ettore",
""
],
[
"Di Gesu",
"Laura",
""
],
[
"Di Lalla",
"Niccolo",
""
],
[
"Donnarumma",
"Immacolata",
""
],
[
"Dovciak",
"Michal",
""
],
[
"Ehlert",
"Steven R.",
""
],
[
"Enoto",
"Teruaki",
""
],
[
"Evangelista",
"Yuri",
""
],
[
"Fabiani",
"Sergio",
""
],
[
"Ferrazzoli",
"Riccardo",
""
],
[
"Garcia",
"Javier A.",
""
],
[
"Gunji",
"Shuichi",
""
],
[
"Hayashida",
"Kiyoshi",
""
],
[
"Iwakiri",
"Wataru",
""
],
[
"Jorstad",
"Svetlana G.",
""
],
[
"Karas",
"Vladimir",
""
],
[
"Kislat",
"Fabian",
""
],
[
"Kitaguchi",
"Takao",
""
],
[
"Kolodziejczak",
"Jeffery J.",
""
],
[
"Latronico",
"Luca",
""
],
[
"Maldera",
"Simone",
""
],
[
"Manfreda",
"Alberto",
""
],
[
"Marin",
"Frederic",
""
],
[
"Marinucci",
"Andrea",
""
],
[
"Marscher",
"Alan P.",
""
],
[
"Marshall",
"Herman L.",
""
],
[
"Massaro",
"Francesco",
""
],
[
"Matt",
"Giorgio",
""
],
[
"Mitsuishi",
"Ikuyuki",
""
],
[
"Mizuno",
"Tsunefumi",
""
],
[
"Muleri",
"Fabio",
""
],
[
"Negro",
"Michela",
""
],
[
"Ng",
"Chi-Yung",
""
],
[
"O'Dell",
"Stephen L.",
""
],
[
"Omodei",
"Nicola",
""
],
[
"Oppedisano",
"Chiara",
""
],
[
"Papitto",
"Alessandro",
""
],
[
"Peirson",
"Abel L.",
""
],
[
"Perri",
"Matteo",
""
],
[
"Pesce-Rollins",
"Melissa",
""
],
[
"Petrucci",
"Pierre-Olivier",
""
],
[
"Pilia",
"Maura",
""
],
[
"Possenti",
"Andrea",
""
],
[
"Puccetti",
"Simonetta",
""
],
[
"Ramsey",
"Brian D.",
""
],
[
"Rankin",
"John",
""
],
[
"Ratheesh",
"Ajay",
""
],
[
"Roberts",
"Oliver J.",
""
],
[
"Romani",
"Roger W.",
""
],
[
"Sgro",
"Carmelo",
""
],
[
"Slane",
"Patrick",
""
],
[
"Soffitta",
"Paolo",
""
],
[
"Spandre",
"Gloria",
""
],
[
"Swartz",
"Douglas A.",
""
],
[
"Tamagawa",
"Toru",
""
],
[
"Tavecchio",
"Fabrizio",
""
],
[
"Taverna",
"Roberto",
""
],
[
"Tawara",
"Yuzuru",
""
],
[
"Tennant",
"Allyn F.",
""
],
[
"Thomas",
"Nicholas E.",
""
],
[
"Tombesi",
"Francesco",
""
],
[
"Trois",
"Alessio",
""
],
[
"Turolla",
"Roberto",
""
],
[
"Vink",
"Jacco",
""
],
[
"Wu",
"Kinwah",
""
],
[
"Xie",
"Fei",
""
]
] |
2405.08177 | Matthew Simpson | Matthew J Simpson and Ruth E Baker | Parameter identifiability, parameter estimation and model prediction for
differential equation models | 22 pages, 6 Figures | null | null | null | stat.ME | http://creativecommons.org/licenses/by/4.0/ | Interpreting data with mathematical models is an important aspect of
real-world applied mathematical modeling. Very often we are interested to
understand the extent to which a particular data set informs and constrains
model parameters. This question is closely related to the concept of parameter
identifiability, and in this article we present a series of computational
exercises to introduce tools that can be used to assess parameter
identifiability, estimate parameters and generate model predictions. Taking a
likelihood-based approach, we show that very similar ideas and algorithms can
be used to deal with a range of different mathematical modelling frameworks.
The exercises and results presented in this article are supported by a suite of
open access codes that can be accessed on GitHub.
| [
{
"version": "v1",
"created": "Mon, 13 May 2024 20:56:05 GMT"
},
{
"version": "v2",
"created": "Tue, 28 May 2024 22:36:47 GMT"
},
{
"version": "v3",
"created": "Thu, 7 Nov 2024 04:00:27 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Simpson",
"Matthew J",
""
],
[
"Baker",
"Ruth E",
""
]
] |
2405.09032 | Jianhua Zhu | Jianhua Zhu, Liangcai Gao and Wenqi Zhao | ICAL: Implicit Character-Aided Learning for Enhanced Handwritten
Mathematical Expression Recognition | ICDAR 2024 Oral Paper | null | 10.1007/978-3-031-70549-6_2 | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Significant progress has been made in the field of handwritten mathematical
expression recognition, while existing encoder-decoder methods are usually
difficult to model global information in $LaTeX$. Therefore, this paper
introduces a novel approach, Implicit Character-Aided Learning (ICAL), to mine
the global expression information and enhance handwritten mathematical
expression recognition. Specifically, we propose the Implicit Character
Construction Module (ICCM) to predict implicit character sequences and use a
Fusion Module to merge the outputs of the ICCM and the decoder, thereby
producing corrected predictions. By modeling and utilizing implicit character
information, ICAL achieves a more accurate and context-aware interpretation of
handwritten mathematical expressions. Experimental results demonstrate that
ICAL notably surpasses the state-of-the-art(SOTA) models, improving the
expression recognition rate (ExpRate) by 2.25\%/1.81\%/1.39\% on the CROHME
2014/2016/2019 datasets respectively, and achieves a remarkable 69.06\% on the
challenging HME100k test set. We make our code available on the GitHub:
https://github.com/qingzhenduyu/ICAL
| [
{
"version": "v1",
"created": "Wed, 15 May 2024 02:03:44 GMT"
},
{
"version": "v2",
"created": "Fri, 16 Aug 2024 07:26:54 GMT"
},
{
"version": "v3",
"created": "Sun, 1 Sep 2024 06:03:39 GMT"
},
{
"version": "v4",
"created": "Thu, 7 Nov 2024 10:06:18 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Zhu",
"Jianhua",
""
],
[
"Gao",
"Liangcai",
""
],
[
"Zhao",
"Wenqi",
""
]
] |
2405.09422 | Daniil Afanasev | D. E. Afanasev and M. O. Katanaev | Was there a Big Bang? | 8 pages, 2 figures; minor corrections | null | null | null | gr-qc hep-th math-ph math.MP | http://creativecommons.org/licenses/by/4.0/ | New one parameter family of exact solutions in General Relativity with a
scalar field is found. The metric is of Liouville type which admits complete
separation of variables in the geodesic Hamilton-Jacobi equation. This solution
exists for the exponential potential for a scalar field and is invariant with
respect to global Lorentz transformations. It describes, in particular,
evolution of the space-time with the naked singularity. Solutions corresponding
to the naked singularity provide accelerating expansion of the homogeneous and
isotropic Universe, and can be smoothly continued along geodesics to infinite
past without Big Bang.
| [
{
"version": "v1",
"created": "Wed, 15 May 2024 15:16:52 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 06:42:53 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Afanasev",
"D. E.",
""
],
[
"Katanaev",
"M. O.",
""
]
] |
2405.09515 | Leonardo Costa Prauchner | Leonardo Prauchner, Eleonir Calegari, Julian Faundez and Sergio
Magalhaes | Interplay between the charge density wave phase and a pseudogap under
antiferromagnetic correlations | 19 pages, 10 figures | null | null | null | cond-mat.str-el | http://creativecommons.org/licenses/by/4.0/ | In this study, we explore the impact of short-range antiferromagnetic
correlations on the charge density wave (CDW) phase in strongly correlated
electron systems exhibiting the pseudogap phenomenon. Our investigation employs
a n-pole approximation to consider the repulsive Coulomb interaction $(U)$ and
antiferromagnetic correlations. Utilizing a two-dimensional Hubbard model for
the Coulomb interaction and a BCS-like model for the CDW order parameter, we
observe that an increase in $U$ enhances antiferromagnetic fluctuations,
resulting in a flattened re-normalized band around the anti-nodal point
$(\pi,0)$. The pseudogap manifests in the band structure and density of states,
prompting an exploration across various $U$ and occupation number values. Our
findings indicate that antiferromagnetic correlations significantly influence
the CDW state, as the Fermi surface is reconstructed within the ordered phase.
Furthermore, we find a Lifhsitz transition inside both the CDW phase and the
normal state, with the latter preceding the onset of the pseudogap.
| [
{
"version": "v1",
"created": "Wed, 15 May 2024 17:08:46 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 16:57:46 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Prauchner",
"Leonardo",
""
],
[
"Calegari",
"Eleonir",
""
],
[
"Faundez",
"Julian",
""
],
[
"Magalhaes",
"Sergio",
""
]
] |
2405.10384 | Musanna Galib | Musanna Galib, Okan K. Orhan, Jian Liu and Mauricio Ponga | Residual Stress Development in Lattice Mismatched Epitaxial Thin Films
via Atomic and Molecular Layer Depositions | null | J. Mech. Phys. Solids. 193 (2024) 105897 | 10.1016/j.jmps.2024.105897 | null | cond-mat.mtrl-sci | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Atomic and molecular layer deposition (ALD/MLD) coatings are promising
solutions for preventing dendrite formation in aqueous and non-aqueous Li/Na/Zn
metal batteries. Notably, alumina and alucone coatings have emerged as highly
effective against dendrite formation in Zn anodes. Despite their demonstrated
efficacy, a comprehensive understanding of their chemo-mechanical impact on
anodes remains elusive. In this study, we take a bottom-up framework to these
coatings on Zn foils, employing an approach that integrates \textit{ab initio}
simulations with continuum theories to elucidate lattice misfit and chemical
bonding. We use this insight to develop a macroscopic model to predict the
epitaxial residual stresses generated during thin-film deposition. Our findings
reveal a robust chemical bonding between the hydroxylated Zn surface and the
thin film. This, in turn, generates large misfit strains that result in
significant interfacial stresses during deposition. These results are then
compared to experiments by measuring the curvature of the coated thin films,
finding good agreement between experiments and theory. This novel understanding
sheds light on the fundamental mechanisms underpinning the development of
chemo-mechanical stresses in thin films, which impact dendrite suppression in
anodes, offering valuable insights for the design of new coatings.
| [
{
"version": "v1",
"created": "Thu, 16 May 2024 18:24:46 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Galib",
"Musanna",
""
],
[
"Orhan",
"Okan K.",
""
],
[
"Liu",
"Jian",
""
],
[
"Ponga",
"Mauricio",
""
]
] |
2405.10649 | Gal Morgenstern | Gal Morgenstern and Tirza Routtenberg | Efficient Recovery of Sparse Graph Signals from Graph Filter Outputs | This work has been submitted to the IEEE for possible publication | null | null | null | eess.SP cs.SY eess.SY math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper investigates the recovery of a node-domain sparse graph signal
from the output of a graph filter. This problem, which is often referred to as
the identification of the source of a diffused sparse graph signal, is seminal
in the field of graph signal processing (GSP). Sparse graph signals can be used
in the modeling of a variety of real-world applications in networks, such as
social, biological, and power systems, and enable various GSP tasks, such as
graph signal reconstruction, blind deconvolution, and sampling. In this paper,
we assume double sparsity of both the graph signal and the graph topology, as
well as a low-order graph filter. We propose three algorithms to reconstruct
the support set of the input sparse graph signal from the graph filter output
samples, leveraging these assumptions and the generalized information criterion
(GIC). First, we describe the graph multiple GIC (GM-GIC) method, which is
based on partitioning the dictionary elements (graph filter matrix columns)
that capture information on the signal into smaller subsets. Then, the local
GICs are computed for each subset and aggregated to make a global decision.
Second, inspired by the well-known branch and bound (BNB) approach, we develop
the graph-based branch and bound GIC (graph-BNB-GIC), and incorporate a new
tractable heuristic bound tailored to the graph and graph filter
characteristics. In addition, we propose the graph-based first order correction
(GFOC) method, which improves existing sparse recovery methods by iteratively
examining potential improvements to the GIC cost function by replacing elements
from the estimated support set with elements from their one-hop neighborhood.
In addition, we investigate the application of our graph-based sparse recovery
methods in blind deconvolution scenarios where the graph filter is unknown.
| [
{
"version": "v1",
"created": "Fri, 17 May 2024 09:23:33 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 13:47:41 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Morgenstern",
"Gal",
""
],
[
"Routtenberg",
"Tirza",
""
]
] |
2405.11488 | Ye Shen | Bin Chen, Yehui Hou, Junyi Li and Ye Shen | Energy Extraction from a Kerr Black Hole via Magnetic Reconnection
within the Plunging Region | null | null | null | null | gr-qc astro-ph.HE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Magnetic reconnection within a highly magnetized plasma has been seen as a
viable mechanism to extract the energy from a rotating black hole, as it can
generate negative energy plasmoids in the ergoregion. For a typical accreting
black hole, the ergoregion is filled with bulk plasma plunging from the
innermost-stable-circular orbit (ISCO). In this study, we present an analytical
study of the energy extraction via magnetic reconnection process in the
plunging region. In contrast to the toroidal plasma, where the magnetic field
cannot be derived from the MHD scheme, the magnetic field in the plunging
plasma was determined by the ideal-MHD condition. We derive the global magnetic
field structure in a fast reconnection model, and we read the expressions for
the energies of plasmoids ejected from the reconnection region, for general
stationary and axisymmetric spacetimes. Then, we demonstrate the behaviors of
ejected energies varying with the reconnection locations in the Kerr spacetime,
and identify the region where a negative-energy plasmoid can be produced. We
find that for a certain magnetization there exists a critical value of the
black hole spin, beyond which the energy extraction can occur, and the energy
extraction is most efficient for the near-extreme black hole. Moreover, we
study the conditions necessary for a plasmoid with positive energy to escape to
the infinity, a crucial requirement for effective energy extractions.
Considering the escaping conditions, we provide the parameter space in the
radius-spin plane in which the energy extraction mechanism is effective.
| [
{
"version": "v1",
"created": "Sun, 19 May 2024 09:15:59 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 12:02:30 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Chen",
"Bin",
""
],
[
"Hou",
"Yehui",
""
],
[
"Li",
"Junyi",
""
],
[
"Shen",
"Ye",
""
]
] |
2405.11600 | Csaba Sandor | Csaba S\'andor and Maciej Zakarczemny | Equal Sum and Product Problem III | The main result of the manuscript has been proved earlier | null | null | null | math.NT | http://creativecommons.org/licenses/by/4.0/ | Denote by $N(n)$ the number of integer solutions $(x_1,\,x_2,\ldots ,x_n)$ of
the equation $x_1+x_2+\ldots+x_n=x_1x_2\cdot\ldots\cdot x_n$ such that $x_1\ge
x_2\ge\ldots\ge x_n\ge 1$, $n \in \mathbb{Z}^+$. The aim of this paper are is
twofold: first we present an asymptotic formula for $\sum\limits_{2\le n\le
x}N(n)$, then we verify that the counting function $N(n)$ takes very large
value compared to its average value.
| [
{
"version": "v1",
"created": "Sun, 19 May 2024 16:07:01 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 14:48:39 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Sándor",
"Csaba",
""
],
[
"Zakarczemny",
"Maciej",
""
]
] |
2405.11663 | Dominique Mouhanna | L. Delzescaux, D. Mouhanna and M. Tissier | Crumpled-to-flat transition of quenched disordered membranes at two-loop
order | 8 pages, 2 figures, published version | Phys. Rev. E 110, 054501 (2024) | 10.1103/PhysRevE.110.054501 | null | cond-mat.dis-nn hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate the effects of quenched elastic disorder on the nature of the
crumpling-to-flat transition of $D$-dimensional polymerized membranes using a
two-loop computation near the upper critical dimension $D_c=4$. While the pure
system undergoes fluctuation-induced first order transitions below $D_c$ and
for an embedding dimension $d<d_{c,pure}\simeq 218.2$, one observes, in
presence of disorder, the emergence of various regions of second order governed
by a disordered stable fixed point for $d<d_{c1}\sim d_{c,pure}$. This opens
the possibility of a new universality class associated with the
crumpling-to-flat transition of disordered membranes in $d=3$
| [
{
"version": "v1",
"created": "Sun, 19 May 2024 20:16:06 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 12:27:02 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Delzescaux",
"L.",
""
],
[
"Mouhanna",
"D.",
""
],
[
"Tissier",
"M.",
""
]
] |
2405.12452 | Junfeng Hu | Junfeng Hu, Xu Liu, Zhencheng Fan, Yifang Yin, Shili Xiang, Savitha
Ramasamy, Roger Zimmermann | Prompt-Based Spatio-Temporal Graph Transfer Learning | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Spatio-temporal graph neural networks have proven efficacy in capturing
complex dependencies for urban computing tasks such as forecasting and kriging.
Yet, their performance is constrained by the reliance on extensive data for
training on a specific task, thereby limiting their adaptability to new urban
domains with varied task demands. Although transfer learning has been proposed
to remedy this problem by leveraging knowledge across domains, the cross-task
generalization still remains under-explored in spatio-temporal graph transfer
learning due to the lack of a unified framework. To bridge the gap, we propose
Spatio-Temporal Graph Prompting (STGP), a prompt-based framework capable of
adapting to multi-diverse tasks in a data-scarce domain. Specifically, we first
unify different tasks into a single template and introduce a task-agnostic
network architecture that aligns with this template. This approach enables
capturing dependencies shared across tasks. Furthermore, we employ learnable
prompts to achieve domain and task transfer in a two-stage prompting pipeline,
facilitating the prompts to effectively capture domain knowledge and
task-specific properties. Our extensive experiments demonstrate that STGP
outperforms state-of-the-art baselines in three tasks-forecasting, kriging, and
extrapolation-achieving an improvement of up to 10.7%.
| [
{
"version": "v1",
"created": "Tue, 21 May 2024 02:06:40 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 06:41:07 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Hu",
"Junfeng",
""
],
[
"Liu",
"Xu",
""
],
[
"Fan",
"Zhencheng",
""
],
[
"Yin",
"Yifang",
""
],
[
"Xiang",
"Shili",
""
],
[
"Ramasamy",
"Savitha",
""
],
[
"Zimmermann",
"Roger",
""
]
] |
2405.12597 | Wen-Fong Ke | Wen-Fong Ke, Johannes H. Meyer | Are zero-symmetric simple nearrings with identity equiprime? | null | null | null | null | math.RA | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We show that there exist zero-symmetric simple nearrings with identity which
are not equiprime, solving a longstanding open problem.
| [
{
"version": "v1",
"created": "Tue, 21 May 2024 08:41:26 GMT"
},
{
"version": "v2",
"created": "Mon, 22 Jul 2024 00:05:34 GMT"
},
{
"version": "v3",
"created": "Sun, 18 Aug 2024 08:33:18 GMT"
},
{
"version": "v4",
"created": "Thu, 7 Nov 2024 01:30:49 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Ke",
"Wen-Fong",
""
],
[
"Meyer",
"Johannes H.",
""
]
] |
2405.14125 | Jingnan Zheng | Jingnan Zheng, Han Wang, An Zhang, Tai D. Nguyen, Jun Sun, Tat-Seng
Chua | ALI-Agent: Assessing LLMs' Alignment with Human Values via Agent-based
Evaluation | null | 2024 Neurips | null | null | cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | Large Language Models (LLMs) can elicit unintended and even harmful content
when misaligned with human values, posing severe risks to users and society. To
mitigate these risks, current evaluation benchmarks predominantly employ
expert-designed contextual scenarios to assess how well LLMs align with human
values. However, the labor-intensive nature of these benchmarks limits their
test scope, hindering their ability to generalize to the extensive variety of
open-world use cases and identify rare but crucial long-tail risks.
Additionally, these static tests fail to adapt to the rapid evolution of LLMs,
making it hard to evaluate timely alignment issues. To address these
challenges, we propose ALI-Agent, an evaluation framework that leverages the
autonomous abilities of LLM-powered agents to conduct in-depth and adaptive
alignment assessments. ALI-Agent operates through two principal stages:
Emulation and Refinement. During the Emulation stage, ALI-Agent automates the
generation of realistic test scenarios. In the Refinement stage, it iteratively
refines the scenarios to probe long-tail risks. Specifically, ALI-Agent
incorporates a memory module to guide test scenario generation, a tool-using
module to reduce human labor in tasks such as evaluating feedback from target
LLMs, and an action module to refine tests. Extensive experiments across three
aspects of human values--stereotypes, morality, and legality--demonstrate that
ALI-Agent, as a general evaluation framework, effectively identifies model
misalignment. Systematic analysis also validates that the generated test
scenarios represent meaningful use cases, as well as integrate enhanced
measures to probe long-tail risks. Our code is available at
https://github.com/SophieZheng998/ALI-Agent.git
| [
{
"version": "v1",
"created": "Thu, 23 May 2024 02:57:42 GMT"
},
{
"version": "v2",
"created": "Fri, 24 May 2024 06:38:49 GMT"
},
{
"version": "v3",
"created": "Thu, 7 Nov 2024 12:07:07 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Zheng",
"Jingnan",
""
],
[
"Wang",
"Han",
""
],
[
"Zhang",
"An",
""
],
[
"Nguyen",
"Tai D.",
""
],
[
"Sun",
"Jun",
""
],
[
"Chua",
"Tat-Seng",
""
]
] |
2405.14758 | Daniel Halpern | Luise Ge, Daniel Halpern, Evi Micha, Ariel D. Procaccia, Itai Shapira,
Yevgeniy Vorobeychik, and Junlin Wu | Axioms for AI Alignment from Human Feedback | null | null | null | null | cs.GT cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the context of reinforcement learning from human feedback (RLHF), the
reward function is generally derived from maximum likelihood estimation of a
random utility model based on pairwise comparisons made by humans. The problem
of learning a reward function is one of preference aggregation that, we argue,
largely falls within the scope of social choice theory. From this perspective,
we can evaluate different aggregation methods via established axioms, examining
whether these methods meet or fail well-known standards. We demonstrate that
both the Bradley-Terry-Luce Model and its broad generalizations fail to meet
basic axioms. In response, we develop novel rules for learning reward functions
with strong axiomatic guarantees. A key innovation from the standpoint of
social choice is that our problem has a linear structure, which greatly
restricts the space of feasible rules and leads to a new paradigm that we call
linear social choice.
| [
{
"version": "v1",
"created": "Thu, 23 May 2024 16:29:29 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 15:40:25 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Ge",
"Luise",
""
],
[
"Halpern",
"Daniel",
""
],
[
"Micha",
"Evi",
""
],
[
"Procaccia",
"Ariel D.",
""
],
[
"Shapira",
"Itai",
""
],
[
"Vorobeychik",
"Yevgeniy",
""
],
[
"Wu",
"Junlin",
""
]
] |
2405.14974 | Zhao Hengyuan | Henry Hengyuan Zhao, Pan Zhou, Difei Gao, Zechen Bai, Mike Zheng Shou | LOVA3: Learning to Visual Question Answering, Asking and Assessment | Accepted by NeurIPS 2024. The code is available at
https://github.com/showlab/LOVA3 | null | null | null | cs.CV cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | Question answering, asking, and assessment are three innate human traits
crucial for understanding the world and acquiring knowledge. By enhancing these
capabilities, humans can more effectively utilize data, leading to better
comprehension and learning outcomes. Current Multimodal Large Language Models
(MLLMs) primarily focus on question answering, often neglecting the full
potential of questioning and assessment skills. Inspired by the human learning
mechanism, we introduce LOVA3, an innovative framework named "Learning tO
Visual question Answering, Asking and Assessment," designed to equip MLLMs with
these additional capabilities. Our approach involves the creation of two
supplementary training tasks GenQA and EvalQA, aiming at fostering the skills
of asking and assessing questions in the context of images. To develop the
questioning ability, we compile a comprehensive set of multimodal foundational
tasks. For assessment, we introduce a new benchmark called EvalQABench,
comprising 64,000 training samples (split evenly between positive and negative
samples) and 5,000 validation and testing samples. We posit that enhancing
MLLMs with the capabilities to answer, ask, and assess questions will enhance
their multimodal comprehension, ultimately improving overall performance. To
validate this hypothesis, we train MLLMs using the LOVA3 framework and evaluate
them on a range of multimodal datasets and benchmarks. Our results demonstrate
consistent performance gains, underscoring the critical role of these
additional tasks in fostering comprehensive intelligence in MLLMs. The code is
available at https://github.com/showlab/LOVA3.
| [
{
"version": "v1",
"created": "Thu, 23 May 2024 18:21:59 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 08:41:03 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Zhao",
"Henry Hengyuan",
""
],
[
"Zhou",
"Pan",
""
],
[
"Gao",
"Difei",
""
],
[
"Bai",
"Zechen",
""
],
[
"Shou",
"Mike Zheng",
""
]
] |
2405.15469 | Joao Yun | Anlaug Amanda Djupvik, Jo\~ao L. Yun, Fernando Comer\'on | Star formation in the high-extinction Planck cold clump PGCC
G120.69+2.66 | 12 pages, 9 figures; accepted revised version | Open Journal of Astrophysics, Vol. 7, 2024 | 10.33232/001c.125627 | null | astro-ph.GA astro-ph.SR | http://creativecommons.org/licenses/by/4.0/ | We investigate the star formation occurring in the Planck Galactic cold clump
PGCC 120.69+2.66. Near-infrared JHKs images and K-band spectroscopy obtained
with NOTCam at the Nordic Optical Telescope complemented with archive data are
used to study the stellar content. In addition, millimetre line CO and CS
spectra were obtained with the Onsala 20 m telescope, and sub-millimetre
continuum SCUBA archive data are used to characterise the host molecular cloud.
We identify a molecular cloud core traced by CO and CS emission at a distance
of 1.1 kpc. In the region studied, we identify 5 submm continuum cores.
Embedded in and around these dense submm cores, we find 38 young stellar
objects, classified as 9 Class I, 8 Class II, and 21 near-IR excess or
variability sources, accompanied by bipolar nebulosities and signs of
protostellar jets. Furthermore, a very bright and reddened source is found
towards this molecular cloud core. Even though its location appears to suggest
its association to the star formation region, its infrared spectral type is
compatible with a red supergiant, hidden behind 36 mag of visual extinction.
| [
{
"version": "v1",
"created": "Fri, 24 May 2024 11:45:04 GMT"
},
{
"version": "v2",
"created": "Wed, 6 Nov 2024 18:26:18 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Djupvik",
"Anlaug Amanda",
""
],
[
"Yun",
"João L.",
""
],
[
"Comerón",
"Fernando",
""
]
] |
2405.15598 | Md Abrar Jahin | Md Abrar Jahin, Asef Shahriar, and Md Al Amin | MCDFN: Supply Chain Demand Forecasting via an Explainable Multi-Channel
Data Fusion Network Model | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Accurate demand forecasting is crucial for optimizing supply chain
management. Traditional methods often fail to capture complex patterns from
seasonal variability and special events. Despite advancements in deep learning,
interpretable forecasting models remain a challenge. To address this, we
introduce the Multi-Channel Data Fusion Network (MCDFN), a hybrid architecture
that integrates Convolutional Neural Networks (CNN), Long Short-Term Memory
networks (LSTM), and Gated Recurrent Units (GRU) to enhance predictive
performance by extracting spatial and temporal features from time series data.
Our comparative benchmarking demonstrates that MCDFN outperforms seven other
deep-learning models, achieving superior metrics: MSE (23.5738), RMSE (4.8553),
MAE (3.9991), and MAPE (20.1575%). Additionally, MCDFN's predictions were
statistically indistinguishable from actual values, confirmed by a paired
t-test with a 5% p-value and a 10-fold cross-validated statistical paired
t-test. We apply explainable AI techniques like ShapTime and Permutation
Feature Importance to enhance interpretability. This research advances demand
forecasting methodologies and offers practical guidelines for integrating MCDFN
into supply chain systems, highlighting future research directions for
scalability and user-friendly deployment.
| [
{
"version": "v1",
"created": "Fri, 24 May 2024 14:30:00 GMT"
},
{
"version": "v2",
"created": "Sun, 23 Jun 2024 17:11:28 GMT"
},
{
"version": "v3",
"created": "Sat, 14 Sep 2024 14:10:52 GMT"
},
{
"version": "v4",
"created": "Thu, 7 Nov 2024 13:26:18 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Jahin",
"Md Abrar",
""
],
[
"Shahriar",
"Asef",
""
],
[
"Amin",
"Md Al",
""
]
] |
2405.16591 | Qijie Wang | Qijie Wang, Guandu Liu, Bin Wang | CapS-Adapter: Caption-based MultiModal Adapter in Zero-Shot
Classification | ACM Multimedia 2024 Poster | null | 10.1145/3664647.3681566 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in vision-language foundational models, such as CLIP, have
demonstrated significant strides in zero-shot classification. However, the
extensive parameterization of models like CLIP necessitates a
resource-intensive fine-tuning process. In response, TIP-Adapter and SuS-X have
introduced training-free methods aimed at bolstering the efficacy of downstream
tasks. While these approaches incorporate support sets to maintain data
distribution consistency between knowledge cache and test sets, they often fall
short in terms of generalization on the test set, particularly when faced with
test data exhibiting substantial distributional variations. In this work, we
present CapS-Adapter, an innovative method that employs a caption-based support
set, effectively harnessing both image and caption features to exceed existing
state-of-the-art techniques in training-free scenarios. CapS-Adapter adeptly
constructs support sets that closely mirror target distributions, utilizing
instance-level distribution features extracted from multimodal large models. By
leveraging CLIP's single and cross-modal strengths, CapS-Adapter enhances
predictive accuracy through the use of multimodal support sets. Our method
achieves outstanding zero-shot classification results across 19 benchmark
datasets, improving accuracy by 2.19\% over the previous leading method. Our
contributions are substantiated through extensive validation on multiple
benchmark datasets, demonstrating superior performance and robust
generalization capabilities. Our code is made publicly available at
https://github.com/WLuLi/CapS-Adapter.
| [
{
"version": "v1",
"created": "Sun, 26 May 2024 14:50:40 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 09:33:40 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Wang",
"Qijie",
""
],
[
"Liu",
"Guandu",
""
],
[
"Wang",
"Bin",
""
]
] |
2405.16598 | Zheng Zhai | Zheng Zhai, Jialu Xu, Mingxin Wu, and Xiaohui Li | Regularized Projection Matrix Approximation with Applications to
Community Detection | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | This paper introduces a regularized projection matrix approximation framework
designed to recover cluster information from the affinity matrix. The model is
formulated as a projection approximation problem, incorporating an entry-wise
penalty function. We investigate three distinct penalty functions, each
specifically tailored to address bounded, positive, and sparse scenarios. To
solve this problem, we propose direct optimization on the Stiefel manifold,
utilizing the Cayley transformation along with the Alternating Direction Method
of Multipliers (ADMM) algorithm. Additionally, we provide a theoretical
analysis that establishes the convergence properties of ADMM, demonstrating
that the convergence point satisfies the KKT conditions of the original
problem. Numerical experiments conducted on both synthetic and real-world
datasets reveal that our regularized projection matrix approximation approach
significantly outperforms state-of-the-art methods in clustering performance.
| [
{
"version": "v1",
"created": "Sun, 26 May 2024 15:18:22 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 13:12:55 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Zhai",
"Zheng",
""
],
[
"Xu",
"Jialu",
""
],
[
"Wu",
"Mingxin",
""
],
[
"Li",
"Xiaohui",
""
]
] |
2405.16907 | Jaewoo Lee | Jaewoo Lee, Sujin Yun, Taeyoung Yun and Jinkyoo Park | GTA: Generative Trajectory Augmentation with Guidance for Offline
Reinforcement Learning | NeurIPS 2024. Previously accepted (Spotlight) to ICLR 2024 Workshop
on Generative Models for Decision Making. Jaewoo Lee and Sujin Yun are equal
contribution authors | null | null | null | cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Offline Reinforcement Learning (Offline RL) presents challenges of learning
effective decision-making policies from static datasets without any online
interactions. Data augmentation techniques, such as noise injection and data
synthesizing, aim to improve Q-function approximation by smoothing the learned
state-action region. However, these methods often fall short of directly
improving the quality of offline datasets, leading to suboptimal results. In
response, we introduce GTA, Generative Trajectory Augmentation, a novel
generative data augmentation approach designed to enrich offline data by
augmenting trajectories to be both high-rewarding and dynamically plausible.
GTA applies a diffusion model within the data augmentation framework. GTA
partially noises original trajectories and then denoises them with
classifier-free guidance via conditioning on amplified return value. Our
results show that GTA, as a general data augmentation strategy, enhances the
performance of widely used offline RL algorithms across various tasks with
unique challenges. Furthermore, we conduct a quality analysis of data augmented
by GTA and demonstrate that GTA improves the quality of the data. Our code is
available at https://github.com/Jaewoopudding/GTA
| [
{
"version": "v1",
"created": "Mon, 27 May 2024 07:55:45 GMT"
},
{
"version": "v2",
"created": "Tue, 28 May 2024 07:00:08 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Jun 2024 05:32:32 GMT"
},
{
"version": "v4",
"created": "Mon, 4 Nov 2024 02:51:13 GMT"
},
{
"version": "v5",
"created": "Thu, 7 Nov 2024 02:24:18 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Lee",
"Jaewoo",
""
],
[
"Yun",
"Sujin",
""
],
[
"Yun",
"Taeyoung",
""
],
[
"Park",
"Jinkyoo",
""
]
] |
2405.17299 | Nikita Tsoy | Nikita Tsoy, Nikola Konstantinov | Simplicity Bias of Two-Layer Networks beyond Linearly Separable Data | ICML 2024, camera-ready version (expanded related work) | null | null | null | stat.ML cs.LG math.OC | http://creativecommons.org/licenses/by/4.0/ | Simplicity bias, the propensity of deep models to over-rely on simple
features, has been identified as a potential reason for limited
out-of-distribution generalization of neural networks (Shah et al., 2020).
Despite the important implications, this phenomenon has been theoretically
confirmed and characterized only under strong dataset assumptions, such as
linear separability (Lyu et al., 2021). In this work, we characterize
simplicity bias for general datasets in the context of two-layer neural
networks initialized with small weights and trained with gradient flow.
Specifically, we prove that in the early training phases, network features
cluster around a few directions that do not depend on the size of the hidden
layer. Furthermore, for datasets with an XOR-like pattern, we precisely
identify the learned features and demonstrate that simplicity bias intensifies
during later training stages. These results indicate that features learned in
the middle stages of training may be more useful for OOD transfer. We support
this hypothesis with experiments on image data.
| [
{
"version": "v1",
"created": "Mon, 27 May 2024 16:00:45 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 13:44:22 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Tsoy",
"Nikita",
""
],
[
"Konstantinov",
"Nikola",
""
]
] |
2405.17382 | Hyunseok Lee | Hyunseok Lee, Jihoon Tack, Jinwoo Shin | ReMoDetect: Reward Models Recognize Aligned LLM's Generations | Published as a conference proceeding for NeurIPS 2024 | null | null | null | cs.LG cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The remarkable capabilities and easy accessibility of large language models
(LLMs) have significantly increased societal risks (e.g., fake news
generation), necessitating the development of LLM-generated text (LGT)
detection methods for safe usage. However, detecting LGTs is challenging due to
the vast number of LLMs, making it impractical to account for each LLM
individually; hence, it is crucial to identify the common characteristics
shared by these models. In this paper, we draw attention to a common feature of
recent powerful LLMs, namely the alignment training, i.e., training LLMs to
generate human-preferable texts. Our key finding is that as these aligned LLMs
are trained to maximize the human preferences, they generate texts with higher
estimated preferences even than human-written texts; thus, such texts are
easily detected by using the reward model (i.e., an LLM trained to model human
preference distribution). Based on this finding, we propose two training
schemes to further improve the detection ability of the reward model, namely
(i) continual preference fine-tuning to make the reward model prefer aligned
LGTs even further and (ii) reward modeling of Human/LLM mixed texts (a
rephrased texts from human-written texts using aligned LLMs), which serves as a
median preference text corpus between LGTs and human-written texts to learn the
decision boundary better. We provide an extensive evaluation by considering six
text domains across twelve aligned LLMs, where our method demonstrates
state-of-the-art results. Code is available at
https://github.com/hyunseoklee-ai/ReMoDetect.
| [
{
"version": "v1",
"created": "Mon, 27 May 2024 17:38:33 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 08:34:27 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Lee",
"Hyunseok",
""
],
[
"Tack",
"Jihoon",
""
],
[
"Shin",
"Jinwoo",
""
]
] |
2405.17767 | Robert Wu | Robert Wu and Vardan Papyan | Linguistic Collapse: Neural Collapse in (Large) Language Models | NeurIPS 2024; 36 pages; 30 figures | null | null | null | cs.LG cs.CL stat.ML | http://creativecommons.org/licenses/by/4.0/ | Neural collapse ($\mathcal{NC}$) is a phenomenon observed in classification
tasks where top-layer representations collapse into their class means, which
become equinorm, equiangular and aligned with the classifiers. These behaviors
-- associated with generalization and robustness -- would manifest under
specific conditions: models are trained towards zero loss, with noise-free
labels belonging to balanced classes, which do not outnumber the model's hidden
dimension. Recent studies have explored $\mathcal{NC}$ in the absence of one or
more of these conditions to extend and capitalize on the associated benefits of
ideal geometries. Language modeling presents a curious frontier, as
\textit{training by token prediction} constitutes a classification task where
none of the conditions exist: the vocabulary is imbalanced and exceeds the
embedding dimension; different tokens might correspond to similar contextual
embeddings; and large language models (LLMs) in particular are typically only
trained for a few epochs. This paper empirically investigates the impact of
scaling the architectures and training of causal language models (CLMs) on
their progression towards $\mathcal{NC}$. We find that $\mathcal{NC}$
properties that develop with scale (and regularization) are linked to
generalization. Moreover, there is evidence of some relationship between
$\mathcal{NC}$ and generalization independent of scale. Our work thereby
underscores the generality of $\mathcal{NC}$ as it extends to the novel and
more challenging setting of language modeling. Downstream, we seek to inspire
further research on the phenomenon to deepen our understanding of LLMs -- and
neural networks at large -- and improve existing architectures based on
$\mathcal{NC}$-related properties. Our code is hosted on GitHub at
https://github.com/rhubarbwu/linguistic-collapse .
| [
{
"version": "v1",
"created": "Tue, 28 May 2024 02:46:11 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 03:16:02 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Wu",
"Robert",
""
],
[
"Papyan",
"Vardan",
""
]
] |
2405.18160 | Christoph Sch\"onle | Christoph Sch\"onle, Marylou Gabri\'e, Tony Leli\`evre, Gabriel Stoltz | Sampling metastable systems using collective variables and
Jarzynski-Crooks paths | Rewritten introduction, added references, small corrections and
clarifications | null | null | null | cond-mat.stat-mech stat.CO | http://creativecommons.org/licenses/by/4.0/ | We consider the problem of sampling a high dimensional multimodal target
probability measure. We assume that a good proposal kernel to move only a
subset of the degrees of freedoms (also known as collective variables) is known
a priori. This proposal kernel can for example be built using normalizing
flows. We show how to extend the move from the collective variable space to the
full space and how to implement an accept-reject step in order to get a
reversible chain with respect to a target probability measure. The
accept-reject step does not require to know the marginal of the original
measure in the collective variable (namely to know the free energy). The
obtained algorithm admits several variants, some of them being very close to
methods which have been proposed previously in the literature. We show how the
obtained acceptance ratio can be expressed in terms of the work which appears
in the Jarzynski-Crooks equality, at least for some variants. Numerical
illustrations demonstrate the efficiency of the approach on various simple test
cases, and allow us to compare the variants of the algorithm.
| [
{
"version": "v1",
"created": "Tue, 28 May 2024 13:22:46 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Jun 2024 07:28:02 GMT"
},
{
"version": "v3",
"created": "Thu, 7 Nov 2024 15:41:11 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Schönle",
"Christoph",
""
],
[
"Gabrié",
"Marylou",
""
],
[
"Lelièvre",
"Tony",
""
],
[
"Stoltz",
"Gabriel",
""
]
] |
2405.18681 | Mauricio Resende | Antonio A. Chaves, Mauricio G.C. Resende, Ricardo M.A. Silva | A random-key GRASP for combinatorial optimization | 24 pages, 8 figures | null | null | null | cs.NE cs.AI math.OC | http://creativecommons.org/licenses/by/4.0/ | This paper proposes a problem-independent GRASP metaheuristic using the
random-key optimizer (RKO) paradigm. GRASP (greedy randomized adaptive search
procedure) is a metaheuristic for combinatorial optimization that repeatedly
applies a semi-greedy construction procedure followed by a local search
procedure. The best solution found over all iterations is returned as the
solution of the GRASP. Continuous GRASP (C-GRASP) is an extension of GRASP for
continuous optimization in the unit hypercube. A random-key optimizer (RKO)
uses a vector of random keys to encode a solution to a combinatorial
optimization problem. It uses a decoder to evaluate a solution encoded by the
vector of random keys. A random-key GRASP is a C-GRASP where points in the unit
hypercube are evaluated employing a decoder. We describe random key GRASP
consisting of a problem-independent component and a problem-dependent decoder.
As a proof of concept, the random-key GRASP is tested on five NP-hard
combinatorial optimization problems: traveling salesman problem, tree of hubs
location problem, Steiner triple covering problem, node capacitated graph
partitioning problem, and job sequencing and tool switching problem.
| [
{
"version": "v1",
"created": "Wed, 29 May 2024 01:07:38 GMT"
},
{
"version": "v2",
"created": "Thu, 30 May 2024 13:06:27 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Chaves",
"Antonio A.",
""
],
[
"Resende",
"Mauricio G. C.",
""
],
[
"Silva",
"Ricardo M. A.",
""
]
] |
2405.18782 | Zihui Wu | Zihui Wu, Yu Sun, Yifan Chen, Bingliang Zhang, Yisong Yue, Katherine
L. Bouman | Principled Probabilistic Imaging using Diffusion Models as Plug-and-Play
Priors | Accepted to NeurIPS 2024 | null | null | null | eess.IV cs.CV stat.ML | http://creativecommons.org/licenses/by/4.0/ | Diffusion models (DMs) have recently shown outstanding capabilities in
modeling complex image distributions, making them expressive image priors for
solving Bayesian inverse problems. However, most existing DM-based methods rely
on approximations in the generative process to be generic to different inverse
problems, leading to inaccurate sample distributions that deviate from the
target posterior defined within the Bayesian framework. To harness the
generative power of DMs while avoiding such approximations, we propose a Markov
chain Monte Carlo algorithm that performs posterior sampling for general
inverse problems by reducing it to sampling the posterior of a Gaussian
denoising problem. Crucially, we leverage a general DM formulation as a unified
interface that allows for rigorously solving the denoising problem with a range
of state-of-the-art DMs. We demonstrate the effectiveness of the proposed
method on six inverse problems (three linear and three nonlinear), including a
real-world black hole imaging problem. Experimental results indicate that our
proposed method offers more accurate reconstructions and posterior estimation
compared to existing DM-based imaging inverse methods.
| [
{
"version": "v1",
"created": "Wed, 29 May 2024 05:42:25 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 01:31:00 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Wu",
"Zihui",
""
],
[
"Sun",
"Yu",
""
],
[
"Chen",
"Yifan",
""
],
[
"Zhang",
"Bingliang",
""
],
[
"Yue",
"Yisong",
""
],
[
"Bouman",
"Katherine L.",
""
]
] |
2405.19279 | Bobby He | Bobby He, Lorenzo Noci, Daniele Paliotta, Imanol Schlag, Thomas
Hofmann | Understanding and Minimising Outlier Features in Neural Network Training | NeurIPS 2024 camera ready | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Outlier Features (OFs) are neurons whose activation magnitudes significantly
exceed the average over a neural network's (NN) width. They are well known to
emerge during standard transformer training and have the undesirable effect of
hindering quantisation in afflicted models. Despite their practical importance,
little is known behind why OFs emerge during training, nor how one can minimise
them.
Our work focuses on the above questions, first identifying several
quantitative metrics, such as the kurtosis over neuron activation norms, to
measure OFs. With these metrics, we study how architectural and optimisation
choices influence OFs, and provide practical insights to minimise OFs during
training. As highlights, we introduce a novel unnormalised transformer block,
the Outlier Protected block, and present a previously unknown benefit of
non-diagonal preconditioning optimisers, finding both approaches to
significantly reduce OFs and improve quantisation without compromising
convergence speed, at scales of up to 7B parameters. Notably, our combination
of OP block and non-diagonal preconditioner (SOAP) achieves 14.87 int8
weight-and-activation perplexity (from 14.71 in standard precision), compared
to 63.4 int8 perplexity (from 16.00) with a default OF-prone combination of
Pre-Norm model and Adam, when quantising OPT-125m models post-training.
Overall, our findings shed new light on our understanding of, our ability to
prevent, and the complexity of this important aspect of NN training dynamics.
| [
{
"version": "v1",
"created": "Wed, 29 May 2024 17:11:28 GMT"
},
{
"version": "v2",
"created": "Wed, 6 Nov 2024 22:45:30 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"He",
"Bobby",
""
],
[
"Noci",
"Lorenzo",
""
],
[
"Paliotta",
"Daniele",
""
],
[
"Schlag",
"Imanol",
""
],
[
"Hofmann",
"Thomas",
""
]
] |
2405.19394 | Henry Davenport | Henry Davenport, Johannes Knolle, Frank Schindler | Interaction-induced crystalline topology of excitons | 11 pages, 5 figures | null | 10.1103/PhysRevLett.133.176601 | null | cond-mat.mes-hall cond-mat.str-el | http://creativecommons.org/licenses/by/4.0/ | We apply the topological theory of symmetry indicators to interaction-induced
exciton band structures in centrosymmetric semiconductors. Crucially, we
distinguish between the topological invariants inherited from the underlying
electron and hole bands, and those that are intrinsic to the exciton
wavefunction itself. Focusing on the latter, we show that there exists a class
of exciton bands for which the maximally-localised exciton Wannier states are
shifted with respect to the electronic Wannier states by a quantised amount; we
call these excitons shift excitons. Our analysis explains how the exciton
spectrum can be topologically nontrivial and sustain exciton edge states in
open boundary conditions even when the underlying noninteracting bands have a
trivial atomic limit. We demonstrate the presence of shift excitons as the
lowest energy neutral excitations of the Su-Schrieffer-Heeger model in its
trivial phase when supplemented by local two-body interactions, and show that
they can be accessed experimentally in local optical conductivity measurements.
| [
{
"version": "v1",
"created": "Wed, 29 May 2024 18:00:00 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Davenport",
"Henry",
""
],
[
"Knolle",
"Johannes",
""
],
[
"Schindler",
"Frank",
""
]
] |
2405.20194 | B.N. Kausik | B.N. Kausik | Occam Gradient Descent | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep learning neural network models must be large enough to adapt to their
problem domain, while small enough to avoid overfitting training data during
gradient descent. To balance these competing demands, overprovisioned deep
learning models such as transformers are trained for a single epoch on large
data sets, and hence inefficient with both computing resources and training
data. In response to these inefficiencies, we exploit learning theory to derive
Occam Gradient Descent, an algorithm that interleaves adaptive reduction of
model size to minimize generalization error, with gradient descent on model
weights to minimize fitting error. In contrast, traditional gradient descent
greedily minimizes fitting error without regard to generalization error. Our
algorithm simultaneously descends the space of weights and topological size of
any neural network without modification. With respect to loss, compute and
model size, our experiments show (a) on image classification benchmarks, linear
and convolutional neural networks trained with Occam Gradient Descent
outperform traditional gradient descent with or without post-train pruning; (b)
on a range of tabular data classification tasks, neural networks trained with
Occam Gradient Descent outperform traditional gradient descent, as well as
Random Forests; (c) on natural language transformers, Occam Gradient Descent
outperforms traditional gradient descent.
| [
{
"version": "v1",
"created": "Thu, 30 May 2024 15:58:22 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Jun 2024 23:51:55 GMT"
},
{
"version": "v3",
"created": "Wed, 17 Jul 2024 14:35:54 GMT"
},
{
"version": "v4",
"created": "Wed, 31 Jul 2024 17:57:33 GMT"
},
{
"version": "v5",
"created": "Thu, 7 Nov 2024 14:25:33 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Kausik",
"B. N.",
""
]
] |
2405.20814 | Pawe{\l} Pietrzycki Dr | Pawe{\l} Pietrzycki, Jan Stochel | Hyperrigidity I: singly generated commutative $C^*$-algebras | null | null | null | null | math.OA math.FA | http://creativecommons.org/licenses/by/4.0/ | In this paper, we study hyperrigidity for singly generated commutative unital
$C^*$-algebras. We look at hyperrigidity from four different points of view.
The first is via operator-moment characterizations of spectral measures, the
subject intensively exploited in quantum theory. The second is based on
dilation theory and the Stone-von Neumann calculus for normal operators. The
third, inspired by a recent result by L. G. Brown, concerns the weak and strong
convergence of sequences of subnormal (or normal) operators. Finally, the
fourth deals with the multiplicativity of unital completely positive maps on
$C^*$-subalgebras generated by single normal elements. The latter is inspired
by Petz's theorem and its generalizations established first by Arveson in the
finite-dimensional case and then by L. G. Brown in general. The above
machinery, together with sophisticated operator inequalities (Kadison's,
Hansen's and Lieb-Ruskai's), makes it possible to identify certain subsets $G$
of the set $\{t^{*m}t^n\colon m, n = 0,1,2,\ldots\}$ that are hyperrigid in
$C^*(t)$, the unital $C^*$-algebra generated by a single normal element $t$.
Examples based on the Choquet boundary are included to confirm the optimality
of our results.
| [
{
"version": "v1",
"created": "Fri, 31 May 2024 14:10:30 GMT"
},
{
"version": "v2",
"created": "Mon, 8 Jul 2024 22:11:07 GMT"
},
{
"version": "v3",
"created": "Wed, 9 Oct 2024 19:01:38 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Pietrzycki",
"Paweł",
""
],
[
"Stochel",
"Jan",
""
]
] |
2406.00599 | Sharmila Duppala | Sharmila Duppala, Juan Luque, John P. Dickerson, Seyed A. Esmaeili | Robust Fair Clustering with Group Membership Uncertainty Sets | null | null | null | null | cs.LG cs.AI cs.CY cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the canonical fair clustering problem where each cluster is
constrained to have close to population-level representation of each group.
Despite significant attention, the salient issue of having incomplete knowledge
about the group membership of each point has been superficially addressed. In
this paper, we consider a setting where the assigned group memberships are
noisy. We introduce a simple noise model that requires a small number of
parameters to be given by the decision maker. We then present an algorithm for
fair clustering with provable \emph{robustness} guarantees. Our framework
enables the decision maker to trade off between the robustness and the
clustering quality. Unlike previous work, our algorithms are backed by
worst-case theoretical guarantees. Finally, we empirically verify the
performance of our algorithm on real world datasets and show its superior
performance over existing baselines.
| [
{
"version": "v1",
"created": "Sun, 2 Jun 2024 03:11:31 GMT"
},
{
"version": "v2",
"created": "Wed, 6 Nov 2024 21:13:21 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Duppala",
"Sharmila",
""
],
[
"Luque",
"Juan",
""
],
[
"Dickerson",
"John P.",
""
],
[
"Esmaeili",
"Seyed A.",
""
]
] |
2406.00691 | Gonzalo Morras | Gonzalo Morras and Juan Garcia-Bellido | Effects of local cosmic inhomogeneities on the gravitational wave event
rate | 8 pages, 5 figures | Phys.Rev.D 110 (2024) 8, 083039 | 10.1103/PhysRevD.110.083039 | IFT-UAM/CSIC-24-064 | astro-ph.CO gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The local universe is highly inhomogeneous and anisotropic. We live in a
relatively sparse region of the Laniakea supercluster at the edge of a large 80
Mpc-wide void. We study the effect of these inhomogeneities on the measured
gravitational wave event rates. In particular, we estimate how the measured
merger rate of compact binaries is biased by the local matter distribution. The
effect of the inhomogeneities on the merger rate is suppressed by the low
angular resolution of gravitational wave detectors coupled with their smoothly
decreasing population-averaged sensitivity with distance. We estimate the
effect on the compact binary coalescence event rate to be at most 6% depending
of the chirp mass of the target binary system and the sensitivity and
orientation of the detectors.
| [
{
"version": "v1",
"created": "Sun, 2 Jun 2024 09:56:40 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 10:11:03 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Morras",
"Gonzalo",
""
],
[
"Garcia-Bellido",
"Juan",
""
]
] |
2406.00922 | Shuyue Stella Li | Shuyue Stella Li, Vidhisha Balachandran, Shangbin Feng, Jonathan S.
Ilgen, Emma Pierson, Pang Wei Koh, Yulia Tsvetkov | MediQ: Question-Asking LLMs and a Benchmark for Reliable Interactive
Clinical Reasoning | 29 pages, 12 figures | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Users typically engage with LLMs interactively, yet most existing benchmarks
evaluate them in a static, single-turn format, posing reliability concerns in
interactive scenarios. We identify a key obstacle towards reliability: LLMs are
trained to answer any question, even with incomplete context or insufficient
knowledge. In this paper, we propose to change the static paradigm to an
interactive one, develop systems that proactively ask questions to gather more
information and respond reliably, and introduce an benchmark - MediQ - to
evaluate question-asking ability in LLMs. MediQ simulates clinical interactions
consisting of a Patient System and an adaptive Expert System; with potentially
incomplete initial information, the Expert refrains from making diagnostic
decisions when unconfident, and instead elicits missing details via follow-up
questions. We provide a pipeline to convert single-turn medical benchmarks into
an interactive format. Our results show that directly prompting
state-of-the-art LLMs to ask questions degrades performance, indicating that
adapting LLMs to proactive information-seeking settings is nontrivial. We
experiment with abstention strategies to better estimate model confidence and
decide when to ask questions, improving diagnostic accuracy by 22.3%; however,
performance still lags compared to an (unrealistic in practice) upper bound
with complete information upfront. Further analyses show improved interactive
performance with filtering irrelevant contexts and reformatting conversations.
Overall, we introduce a novel problem towards LLM reliability, an interactive
MediQ benchmark and a novel question-asking system, and highlight directions to
extend LLMs' information-seeking abilities in critical domains.
| [
{
"version": "v1",
"created": "Mon, 3 Jun 2024 01:32:52 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Jun 2024 13:55:05 GMT"
},
{
"version": "v3",
"created": "Thu, 7 Nov 2024 18:59:30 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Li",
"Shuyue Stella",
""
],
[
"Balachandran",
"Vidhisha",
""
],
[
"Feng",
"Shangbin",
""
],
[
"Ilgen",
"Jonathan S.",
""
],
[
"Pierson",
"Emma",
""
],
[
"Koh",
"Pang Wei",
""
],
[
"Tsvetkov",
"Yulia",
""
]
] |
2406.00925 | Norton Lee | Kimyeong Lee and Norton Lee | Dimers for Type D Relativistic Toda Model | 25+6 pages, 14 figures, add citation | JHEP09(2024)198 | 10.1007/JHEP09(2024)198 | KIAS-P24038, CGP24008 | hep-th | http://creativecommons.org/licenses/by/4.0/ | We construct dimer graphs for type D relativistic Toda models by introducing
impurities to the $Y^{2N,0}$ square dimer graphs. By properly placing the
impurities and change of canonical variables assigned to the 1-loops on the
dimer graph, we introduce the "folding" of the graphs and get the type D
relativistic Toda lattice Hamiltonian and monodromy matrix.
| [
{
"version": "v1",
"created": "Mon, 3 Jun 2024 01:47:38 GMT"
},
{
"version": "v2",
"created": "Sun, 23 Jun 2024 23:25:32 GMT"
},
{
"version": "v3",
"created": "Mon, 2 Sep 2024 04:19:24 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Lee",
"Kimyeong",
""
],
[
"Lee",
"Norton",
""
]
] |
2406.01494 | Ba-Hien Tran | Markus Heinonen, Ba-Hien Tran, Michael Kampffmeyer, Maurizio Filippone | Robust Classification by Coupling Data Mollification with Label
Smoothing | Under review | null | null | null | cs.CV cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | Introducing training-time augmentations is a key technique to enhance
generalization and prepare deep neural networks against test-time corruptions.
Inspired by the success of generative diffusion models, we propose a novel
approach of coupling data mollification, in the form of image noising and
blurring, with label smoothing to align predicted label confidences with image
degradation. The method is simple to implement, introduces negligible
overheads, and can be combined with existing augmentations. We demonstrate
improved robustness and uncertainty quantification on the corrupted image
benchmarks of the CIFAR and TinyImageNet datasets.
| [
{
"version": "v1",
"created": "Mon, 3 Jun 2024 16:21:29 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 09:23:34 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Heinonen",
"Markus",
""
],
[
"Tran",
"Ba-Hien",
""
],
[
"Kampffmeyer",
"Michael",
""
],
[
"Filippone",
"Maurizio",
""
]
] |
2406.02817 | Arvind Rajaraman | Arvind Rajaraman | Instabilities of Gauged Q-Balls | 12 pages | null | null | null | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present the first analytical calculation that shows that perturbations
with angular dependence can lead to an instability in gauged Q-balls. We find
an explicit condition on the parameters for the Q-ball to become unstable. We
compare our predictions to the numerical calculation in Kinach et al., and show
agreement, including a correct prediction of the instability/stability of the
two parameter points analyzed in that paper.
| [
{
"version": "v1",
"created": "Tue, 4 Jun 2024 23:33:08 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Jun 2024 18:06:56 GMT"
},
{
"version": "v3",
"created": "Wed, 6 Nov 2024 22:05:25 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Rajaraman",
"Arvind",
""
]
] |
2406.02857 | Chandra Shekhar Saraf | Chandra Shekhar Saraf, Pawel Bielewicz and Michal Chodorowski | Effect of redshift bin mismatch on cross correlation between DESI Legacy
Imaging Survey and Planck CMB lensing potential | 20 pages, 17 figures, 4 tables, published in Astronomy and
Astrophysics | Astronomy & Astrophysics, Volume 690, October 2024, id.A338, 20 pp | 10.1051/0004-6361/202450749 | null | astro-ph.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the importance of precise modelling of the photometric redshift
error distributions on the estimation of parameters from cross correlation
measurements and present a working example of the scattering matrix formalism
to correct for the redshift bin mismatch of objects in tomographic cross
correlation analysis. We measured the angular galaxy auto-power spectrum and
cross-power spectrum in four tomographic bins with redshift intervals $z =
[0.0,0.3,0.45,0.6,0.8]$ from the cross correlation of Planck Cosmic Microwave
Background lensing potential and photometric galaxy catalogue from the Dark
Energy Spectroscopic Instrument Legacy Imaging Survey Data Release 8. We
estimated galaxy linear bias and amplitude of cross correlation using maximum
likelihood estimation to put constraints on the $\sigma_{8}$ parameter. We show
that the modified Lorentzian function used to fit the photometric redshift
error distribution performs well only near the peaks of the distribution. We
adopt a sum of Gaussians model to capture the broad tails of the error
distribution. Our sum of Gaussians model yields $\sim 2-5\,\sigma$ smaller
values of cross correlation amplitude compared to the $\Lambda$CDM
expectations. We compute the $\sigma_{8}$ parameter after correcting for the
redshift bin mismatch of objects following the scattering matrix approach. The
$\sigma_{8}$ parameter becomes consistent with $\Lambda$CDM model in the last
tomographic bin but shows $\sim 1-3\,\sigma$ tension in other redshift bins.
| [
{
"version": "v1",
"created": "Wed, 5 Jun 2024 02:14:34 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 06:12:12 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Saraf",
"Chandra Shekhar",
""
],
[
"Bielewicz",
"Pawel",
""
],
[
"Chodorowski",
"Michal",
""
]
] |
2406.02880 | Jaiying Shi | Dong Zhao and Jiaying Shi and Wenjun Li and Shudong Wang and Shenghui
Xu and Zhaoming Pan | Controllable Talking Face Generation by Implicit Facial Keypoints
Editing | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Audio-driven talking face generation has garnered significant interest within
the domain of digital human research. Existing methods are encumbered by
intricate model architectures that are intricately dependent on each other,
complicating the process of re-editing image or video inputs. In this work, we
present ControlTalk, a talking face generation method to control face
expression deformation based on driven audio, which can construct the head pose
and facial expression including lip motion for both single image or sequential
video inputs in a unified manner. By utilizing a pre-trained video synthesis
renderer and proposing the lightweight adaptation, ControlTalk achieves precise
and naturalistic lip synchronization while enabling quantitative control over
mouth opening shape. Our experiments show that our method is superior to
state-of-the-art performance on widely used benchmarks, including HDTF and
MEAD. The parameterized adaptation demonstrates remarkable generalization
capabilities, effectively handling expression deformation across same-ID and
cross-ID scenarios, and extending its utility to out-of-domain portraits,
regardless of languages. Code is available at
https://github.com/NetEase-Media/ControlTalk.
| [
{
"version": "v1",
"created": "Wed, 5 Jun 2024 02:54:46 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 02:26:49 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Zhao",
"Dong",
""
],
[
"Shi",
"Jiaying",
""
],
[
"Li",
"Wenjun",
""
],
[
"Wang",
"Shudong",
""
],
[
"Xu",
"Shenghui",
""
],
[
"Pan",
"Zhaoming",
""
]
] |
2406.03415 | Matthew Brehmer | Matthew Brehmer, Margaret Drouhard, Arjun Srinivasan | RemixTape: Enriching Narratives about Metrics with Semantic Alignment
and Contextual Recommendation | SUPPLEMENTAL MATERIAL. The scenario video (Sec. 4.4):
https://youtu.be/9h6aHvqg9vA; the tutorial video (Sec. 5.2):
https://youtu.be/uZPZ5-oiqZk | null | null | null | cs.HC | http://creativecommons.org/licenses/by/4.0/ | The temporal dynamics of quantitative metrics or key performance indicators
(KPIs) are central to decision-making in enterprise organizations. Recently,
major business intelligence providers have introduced new infrastructure for
defining, sharing, and monitoring metric values. However, these values are
often presented in isolation and appropriate context is seldom externalized. In
this design study, we present REMIXTAPE, an application for constructing
structured narratives around metrics. With design imperatives grounded in prior
work and a formative interview study, REMIXTAPE provides a hierarchical canvas
for collecting and coordinating sequences of line chart representations of
metrics, along with the ability to externalize situational context around them.
REMIXTAPE includes affordances to semantically align and annotate juxtaposed
charts and text, as well as recommendations of complementary charts based on
metrics already present on the canvas. We evaluated REMIXTAPE in a study in
which six enterprise data professionals reproduced and extended partial
narratives. They appreciated REMIXTAPE as a novel alternative to dashboards,
galleries, and slide presentations for supporting conversations about metrics.
We conclude with a reflection on our design choices and process, with a call to
define a conceptual foundation for remixing in the context of visualization.
| [
{
"version": "v1",
"created": "Wed, 5 Jun 2024 16:11:15 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 18:52:30 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Brehmer",
"Matthew",
""
],
[
"Drouhard",
"Margaret",
""
],
[
"Srinivasan",
"Arjun",
""
]
] |
2406.03554 | Gang Yao | Yu Liao, Yueqiao Qu, Zuo Li, Yu Chen, Liang Liu, Jun-Zhong Wang, and
Gang Yao | Magnetic ground state and strain-mediated chiral-like atomic distortions
behavior in two-dimensional rectangular spin lattice | 18 pages, 7 figures | Phys. Rev. B 110, 085432 (2024) | 10.1103/PhysRevB.110.085432 | null | cond-mat.mtrl-sci | http://creativecommons.org/publicdomain/zero/1.0/ | Due to the large perpendicular magnetic anisotropy originating from
spin-orbit coupling, magnetoelastic coupling is generally reported in
easy-plane magnets with rectangular lattice where the easy magnetization is
coupled with the lattice direction, while the acquisition of a novel coupling,
beyond the easy-plane ferromagnets, in two-dimensional (2D) materials remains
unknown. Here, by employing the density functional theory calculations, we
demonstrate this feasibility with the discovery of long-range ferromagnetic
ordering and elastic strain-mediated chiral-like atomic distortions behavior in
a newly tetragonal As-Fe-As trilayer (t-FeAs monolayer), which shows large
perpendicular magnetic anisotropy, robust ferromagnetic ordering, and in-plane
ferroelasticity. We firstly point out that obvious limits exist when using the
four magnetic configurations to determine the magnetic ground state for a
rectangular spin lattice even if more exchange interaction parameters are
included. A four-state mapping analysis is carefully examined for t-FeAs, where
the calculated Curie temperature, Tc, is 435 K, which is higher than most
reported 2D magnets, and can be further tuned by appropriate strains.
Intriguingly, the chiral-like atomic distortion behavior of the Fe sub-layer is
scanning tunneling microscopy characterizable, which can switch the
magnetization axis between the out-of-plane and in-plane direction. This
unusual finding of ferroelastic manipulation of both the atomic displacement
and spin properties makes t-FeAs a promising candidate for future spintronics
and also provides the possibility for exploring unprecedented coupling physics.
| [
{
"version": "v1",
"created": "Wed, 5 Jun 2024 18:08:25 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Liao",
"Yu",
""
],
[
"Qu",
"Yueqiao",
""
],
[
"Li",
"Zuo",
""
],
[
"Chen",
"Yu",
""
],
[
"Liu",
"Liang",
""
],
[
"Wang",
"Jun-Zhong",
""
],
[
"Yao",
"Gang",
""
]
] |
2406.03619 | Ben Shaw | Ben Shaw, Abram Magner, Kevin R. Moon | Symmetry Discovery Beyond Affine Transformations | null | null | null | null | cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | Symmetry detection can improve various machine learning tasks. In the context
of continuous symmetry detection, current state of the art experiments are
limited to detecting affine transformations. Under the manifold assumption, we
outline a framework for discovering continuous symmetry in data beyond the
affine transformation group. We also provide a similar framework for
discovering discrete symmetry. We experimentally compare our method to an
existing method known as LieGAN and show that our method is competitive at
detecting affine symmetries for large sample sizes and superior than LieGAN for
small sample sizes. We also show our method is able to detect continuous
symmetries beyond the affine group and is generally more computationally
efficient than LieGAN.
| [
{
"version": "v1",
"created": "Wed, 5 Jun 2024 20:38:30 GMT"
},
{
"version": "v2",
"created": "Wed, 6 Nov 2024 22:25:43 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Shaw",
"Ben",
""
],
[
"Magner",
"Abram",
""
],
[
"Moon",
"Kevin R.",
""
]
] |
2406.05768 | Qingsong Xie | Qingsong Xie, Zhenyi Liao, Zhijie Deng, Chen chen, Haonan Lu | TLCM: Training-efficient Latent Consistency Model for Image Generation
with 2-8 Steps | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Distilling latent diffusion models (LDMs) into ones that are fast to sample
from is attracting growing research interest. However, the majority of existing
methods face two critical challenges: (1) They hinge on long training using a
huge volume of real data. (2) They routinely lead to quality degradation for
generation, especially in text-image alignment. This paper proposes a novel
training-efficient Latent Consistency Model (TLCM) to overcome these
challenges. Our method first accelerates LDMs via data-free multistep latent
consistency distillation (MLCD), and then data-free latent consistency
distillation is proposed to efficiently guarantee the inter-segment consistency
in MLCD. Furthermore, we introduce bags of techniques, e.g., distribution
matching, adversarial learning, and preference learning, to enhance TLCM's
performance at few-step inference without any real data. TLCM demonstrates a
high level of flexibility by enabling adjustment of sampling steps within the
range of 2 to 8 while still producing competitive outputs compared to full-step
approaches. Notably, TLCM enjoys the data-free merit by employing synthetic
data from the teacher for distillation. With just 70 training hours on an A100
GPU, a 3-step TLCM distilled from SDXL achieves an impressive CLIP Score of
33.68 and an Aesthetic Score of 5.97 on the MSCOCO-2017 5K benchmark,
surpassing various accelerated models and even outperforming the teacher model
in human preference metrics. We also demonstrate the versatility of TLCMs in
applications including image style transfer, controllable generation, and
Chinese-to-image generation.
| [
{
"version": "v1",
"created": "Sun, 9 Jun 2024 12:55:50 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Jun 2024 06:22:53 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Jun 2024 02:57:00 GMT"
},
{
"version": "v4",
"created": "Wed, 30 Oct 2024 06:49:52 GMT"
},
{
"version": "v5",
"created": "Thu, 31 Oct 2024 02:16:04 GMT"
},
{
"version": "v6",
"created": "Thu, 7 Nov 2024 01:29:26 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Xie",
"Qingsong",
""
],
[
"Liao",
"Zhenyi",
""
],
[
"Deng",
"Zhijie",
""
],
[
"chen",
"Chen",
""
],
[
"Lu",
"Haonan",
""
]
] |
2406.06005 | Chong Zhang | Chong Zhang, Wenli Xiao, Tairan He, Guanya Shi | WoCoCo: Learning Whole-Body Humanoid Control with Sequential Contacts | Website, Code, and Videos: https://lecar-lab.github.io/wococo/ | null | null | null | cs.RO cs.GR cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Humanoid activities involving sequential contacts are crucial for complex
robotic interactions and operations in the real world and are traditionally
solved by model-based motion planning, which is time-consuming and often relies
on simplified dynamics models. Although model-free reinforcement learning (RL)
has become a powerful tool for versatile and robust whole-body humanoid
control, it still requires tedious task-specific tuning and state machine
design and suffers from long-horizon exploration issues in tasks involving
contact sequences. In this work, we propose WoCoCo (Whole-Body Control with
Sequential Contacts), a unified framework to learn whole-body humanoid control
with sequential contacts by naturally decomposing the tasks into separate
contact stages. Such decomposition facilitates simple and general policy
learning pipelines through task-agnostic reward and sim-to-real designs,
requiring only one or two task-related terms to be specified for each task. We
demonstrated that end-to-end RL-based controllers trained with WoCoCo enable
four challenging whole-body humanoid tasks involving diverse contact sequences
in the real world without any motion priors: 1) versatile parkour jumping, 2)
box loco-manipulation, 3) dynamic clap-and-tap dancing, and 4) cliffside
climbing. We further show that WoCoCo is a general framework beyond humanoid by
applying it in 22-DoF dinosaur robot loco-manipulation tasks.
| [
{
"version": "v1",
"created": "Mon, 10 Jun 2024 04:00:55 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 13:59:05 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Zhang",
"Chong",
""
],
[
"Xiao",
"Wenli",
""
],
[
"He",
"Tairan",
""
],
[
"Shi",
"Guanya",
""
]
] |
2406.07159 | Wei-Han Li | Wei-Han Li and Abbas Ali Saberi | Eigenstate plateau transition and equilibration in 1D quantum lattice
models | 10 pages, 7 figures | null | 10.1103/PhysRevB.110.184104 | null | cond-mat.quant-gas | http://creativecommons.org/licenses/by/4.0/ | We report on a remarkable spectral phenomenon in a generic type of quantum
lattice gas model. As the interaction strength increases, eigenstates
spontaneously reorganize and lead to plateaus of the interaction energy, with
gaps opening akin to continuous phase transitions. Perturbation theory
identifies a hidden structure underlying eigenstates within each plateau,
resulting in a statistical shift in the wavefunction amplitudes described by
extreme value theory. The structured eigenstates manifest themselves naturally
in far-from-equilibrium dynamics proceeding through multiple universal stages.
Our findings reveal a profound connection between emergent properties in
high-energy states and out-of-equilibrium dynamics, providing insights into the
impact of interactions across the entire energy spectrum. The results are
directly relevant to experiments probing equilibration in quantum spin and
lattice gases.
| [
{
"version": "v1",
"created": "Tue, 11 Jun 2024 11:06:01 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 15:08:52 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Li",
"Wei-Han",
""
],
[
"Saberi",
"Abbas Ali",
""
]
] |
2406.07765 | Yaroslav Golubev | Agnia Sergeyuk, Yaroslav Golubev, Timofey Bryksin, Iftekhar Ahmed | Using AI-Based Coding Assistants in Practice: State of Affairs,
Perceptions, and Ways Forward | Published in Information and Software Technology. 32 pages, 4 figures | null | 10.1016/j.infsof.2024.107610 | null | cs.SE cs.AI cs.CY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Context. The last several years saw the emergence of AI assistants for code -
multi-purpose AI-based helpers in software engineering. As they become
omnipresent in all aspects of software development, it becomes critical to
understand their usage patterns.
Objective. We aim to better understand how specifically developers are using
AI assistants, why they are not using them in certain parts of their
development workflow, and what needs to be improved in the future.
Methods. In this work, we carried out a large-scale survey aimed at how AI
assistants are used, focusing on specific software development activities and
stages. We collected opinions of 481 programmers on five broad activities: (a)
implementing new features, (b) writing tests, (c) bug triaging, (d)
refactoring, and (e) writing natural-language artifacts, as well as their
individual stages.
Results. Our results provide a novel comparison of different stages where AI
assistants are used that is both comprehensive and detailed. It highlights
specific activities that developers find less enjoyable and want to delegate to
an AI assistant, e.g., writing tests and natural-language artifacts. We also
determine more granular stages where AI assistants are used, such as generating
tests and generating docstrings, as well as less studied parts of the workflow,
such as generating test data. Among the reasons for not using assistants, there
are general aspects like trust and company policies, as well as more concrete
issues like the lack of project-size context, which can be the focus of the
future research.
Conclusion. The provided analysis highlights stages of software development
that developers want to delegate and that are already popular for using AI
assistants, which can be a good focus for features aimed to help developers
right now. The main reasons for not using AI assistants can serve as a
guideline for future work.
| [
{
"version": "v1",
"created": "Tue, 11 Jun 2024 23:10:43 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 11:09:24 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Sergeyuk",
"Agnia",
""
],
[
"Golubev",
"Yaroslav",
""
],
[
"Bryksin",
"Timofey",
""
],
[
"Ahmed",
"Iftekhar",
""
]
] |
2406.08387 | Gregory Schehr | Marco Biroli, Yannick Feld, Alexander K. Hartmann, Satya N. Majumdar,
Gregory Schehr | Resetting by rescaling: exact results for a diffusing particle in
one-dimension | 19 pages, 8 figures | Phys. Rev. E 110(4), 044142 (2024) | 10.1103/PhysRevE.110.044142 | null | cond-mat.stat-mech | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we study a simple model of a diffusive particle on a line,
undergoing a stochastic resetting with rate $r$, via rescaling its current
position by a factor $a$, which can be either positive or negative. For
$|a|<1$, the position distribution becomes stationary at long times and we
compute this limiting distribution exactly for all $|a|<1$. This symmetric
distribution has a Gaussian shape near its peak at $x=0$, but decays
exponentially for large $|x|$. We also studied the mean first-passage time
(MFPT) $T(0)$ to a target located at a distance $L$ from the initial position
(the origin) of the particle. As a function of the initial position $x$, the
MFPT $T(x)$ satisfies a nonlocal second order differential equation and we have
solved it explicitly for $0 \leq a < 1$. For $-1<a\leq 0$, we also solved it
analytically but up to a constant factor $\kappa$ whose value can be determined
independently from numerical simulations. Our results show that, for all
$-1<a<1$, the MFPT $T(0)$ (starting from the origin) shows a minimum at
$r=r^*(a)$. However, the optimised MFPT $T_{\rm opt}(a)$ turns out to be a
monotonically increasing function of $a$ for $-1<a<1$. This demonstrates that,
compared to the standard resetting to the origin ($a=0$), while the positive
rescaling is not beneficial for the search of a target, the negative rescaling
is. Thus resetting via rescaling followed by a reflection around the origin
expedites the search of a target in one dimension.
| [
{
"version": "v1",
"created": "Wed, 12 Jun 2024 16:37:45 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Biroli",
"Marco",
""
],
[
"Feld",
"Yannick",
""
],
[
"Hartmann",
"Alexander K.",
""
],
[
"Majumdar",
"Satya N.",
""
],
[
"Schehr",
"Gregory",
""
]
] |
2406.08506 | Micha{\l} Koziarski | Micha{\l} Koziarski, Andrei Rekesh, Dmytro Shevchuk, Almer van der
Sloot, Piotr Gai\'nski, Yoshua Bengio, Cheng-Hao Liu, Mike Tyers, Robert A.
Batey | RGFN: Synthesizable Molecular Generation Using GFlowNets | null | null | null | null | physics.chem-ph cs.LG q-bio.BM | http://creativecommons.org/licenses/by/4.0/ | Generative models hold great promise for small molecule discovery,
significantly increasing the size of search space compared to traditional in
silico screening libraries. However, most existing machine learning methods for
small molecule generation suffer from poor synthesizability of candidate
compounds, making experimental validation difficult. In this paper we propose
Reaction-GFlowNet (RGFN), an extension of the GFlowNet framework that operates
directly in the space of chemical reactions, thereby allowing out-of-the-box
synthesizability while maintaining comparable quality of generated candidates.
We demonstrate that with the proposed set of reactions and building blocks, it
is possible to obtain a search space of molecules orders of magnitude larger
than existing screening libraries coupled with low cost of synthesis. We also
show that the approach scales to very large fragment libraries, further
increasing the number of potential molecules. We demonstrate the effectiveness
of the proposed approach across a range of oracle models, including pretrained
proxy models and GPU-accelerated docking.
| [
{
"version": "v1",
"created": "Sat, 1 Jun 2024 13:11:11 GMT"
},
{
"version": "v2",
"created": "Wed, 6 Nov 2024 21:25:19 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Koziarski",
"Michał",
""
],
[
"Rekesh",
"Andrei",
""
],
[
"Shevchuk",
"Dmytro",
""
],
[
"van der Sloot",
"Almer",
""
],
[
"Gaiński",
"Piotr",
""
],
[
"Bengio",
"Yoshua",
""
],
[
"Liu",
"Cheng-Hao",
""
],
[
"Tyers",
"Mike",
""
],
[
"Batey",
"Robert A.",
""
]
] |
2406.08578 | Itzhak Goldman | Itzhak Goldman | Evidence for large scale compressible turbulence in the ism of CSWA13, a
star-Forming Lensed Galaxy at z = 1.87 with outflowing wind | 8 pages, 16 figures, accepted for publication in the open journal of
astrophysics | The Open Journal of Astrophysics, Vol. 7, November 6 (2024) | 10.33232/001c.125516 | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | Recently, Keerthi Vasan G. et al.(2024 presented spatially resolved
observations of a wind outflow in CSWA13, a gravitationally lensed Star-Forming
galaxy at z = 1.87. The gravitational lensing allowed for a substantially
improved spatial and kinematic resolution of the wind and of the nebular gas.
In this paper we take advantage of the resolved data to test for the existence
of turbulence and to study its nature.We obtained the autocorrelation functions
of the two velocity fields, and derived the spatial structure functions of the
residual nebular and wind velocities along the major axis of the galaxy. The
structure functions, of both velocity fields, reveal the existence of an
underlying Burgers power spectrum scaling. This scaling suggests the existence
of supersonic compressible turbulence. The autocorrelation functions exhibit
correlations over scales comparable to the total extent of the velocity fields.
Thus, the turbulence is a large scale one. The turbulent timescale
corresponding to the largest scale is about 200 Myr, an order of magnitude
larger than the estimated age of the wind and of the young stars. This implies
that the turbulence in the ism formed well before the wind and the young stars.
Given the large spatial scale of the turbulence, it is plausible that the
source of the turbulence is a large scale one e.g. a merger or tidal event that
triggered the formation of molecular clouds, in the cores of which, the young
stars formed. A steepening of the structure functions on the smaller scales
provides an estimate for the effective depth along the line of sight of the
turbulent layer. The latter turns out to be about 2kpc
| [
{
"version": "v1",
"created": "Wed, 12 Jun 2024 18:31:08 GMT"
},
{
"version": "v2",
"created": "Mon, 4 Nov 2024 20:19:41 GMT"
},
{
"version": "v3",
"created": "Thu, 7 Nov 2024 07:55:52 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Goldman",
"Itzhak",
""
]
] |
2406.08711 | Robin Bowers | Robin Bowers, Bo Waggoner | Matching with Nested and Bundled Pandora Boxes | null | null | 10.1007/978-3-031-48974-7_4 | null | cs.DS cs.GT | http://creativecommons.org/licenses/by/4.0/ | We consider max-weighted matching with costs for learning the weights,
modeled as a "Pandora's Box" on each endpoint of an edge. Each vertex has an
initially-unknown value for being matched to a neighbor, and an algorithm must
pay some cost to observe this value. The goal is to maximize the total matched
value minus costs. Our model is inspired by two-sided settings, such as
matching employees to employers. Importantly for such settings, we allow for
negative values which cause existing approaches to fail.
We first prove upper bounds for algorithms in two natural classes. Any
algorithm that "bundles" the two Pandora boxes incident to an edge is an
$o(1)$-approximation. Likewise, any "vertex-based" algorithm, which uses
properties of the separate Pandora's boxes but does not consider the
interaction of their value distributions, is an $o(1)$-approximation. Instead,
we utilize Pandora's Nested-Box Problem, i.e. multiple stages of inspection. We
give a self-contained, fully constructive optimal solution to the nested-boxes
problem, which may have structural observations of interest compared to prior
work. By interpreting each edge as a nested box, we leverage this solution to
obtain a constant-factor approximation algorithm. Finally, we show any
"edge-based" algorithm, which considers the interactions of values along an
edge but not with the rest of the graph, is also an $o(1)$-approximation.
| [
{
"version": "v1",
"created": "Thu, 13 Jun 2024 00:25:36 GMT"
},
{
"version": "v2",
"created": "Wed, 6 Nov 2024 19:22:50 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Bowers",
"Robin",
""
],
[
"Waggoner",
"Bo",
""
]
] |
2406.09010 | Vivekananda Roy | Vivekananda Roy | A geometric approach to informed MCMC sampling | null | null | null | null | stat.ME stat.CO | http://creativecommons.org/licenses/by/4.0/ | A Riemannian geometric framework for Markov chain Monte Carlo (MCMC) is
developed where using the Fisher-Rao metric on the manifold of probability
density functions (pdfs), informed proposal densities for Metropolis-Hastings
(MH) algorithms are constructed. We exploit the square-root representation of
pdfs under which the Fisher-Rao metric boils down to the standard $L^2$ metric
on the positive orthant of the unit hypersphere. The square-root representation
allows us to easily compute the geodesic distance between densities, resulting
in a straightforward implementation of the proposed geometric MCMC methodology.
Unlike the random walk MH that blindly proposes a candidate state using no
information about the target, the geometric MH algorithms move an uninformed
base density (e.g., a random walk proposal density) towards different
global/local approximations of the target density, allowing effective
exploration of the distribution simultaneously at different granular levels of
the state space. We compare the proposed geometric MH algorithm with other MCMC
algorithms for various Markov chain orderings, namely the covariance,
efficiency, Peskun, and spectral gap orderings. The superior performance of the
geometric algorithms over other MH algorithms like the random walk Metropolis,
independent MH, and variants of Metropolis adjusted Langevin algorithms is
demonstrated in the context of various multimodal, nonlinear, and high
dimensional examples. In particular, we use extensive simulation and real data
applications to compare these algorithms for analyzing mixture models, logistic
regression models, spatial generalized linear mixed models and ultra-high
dimensional Bayesian variable selection models. A publicly available R package
accompanies the article.
| [
{
"version": "v1",
"created": "Thu, 13 Jun 2024 11:30:29 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 09:24:15 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Roy",
"Vivekananda",
""
]
] |
2406.09014 | Tomas Kulvicius | Tomas Kulvicius, Dajie Zhang, Luise Poustka, Sven B\"olte, Lennart
Jahn, Sarah Fl\"ugge, Marc Kraft, Markus Zweckstetter, Karin Nielsen-Saines,
Florentin W\"org\"otter, and Peter B Marschik | Deep learning empowered sensor fusion boosts infant movement
classification | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | To assess the integrity of the developing nervous system, the Prechtl general
movement assessment (GMA) is recognized for its clinical value in diagnosing
neurological impairments in early infancy. GMA has been increasingly augmented
through machine learning approaches intending to scale-up its application,
circumvent costs in the training of human assessors and further standardize
classification of spontaneous motor patterns. Available deep learning tools,
all of which are based on single sensor modalities, are however still
considerably inferior to that of well-trained human assessors. These approaches
are hardly comparable as all models are designed, trained and evaluated on
proprietary/silo-data sets. With this study we propose a sensor fusion approach
for assessing fidgety movements (FMs). FMs were recorded from 51 typically
developing participants. We compared three different sensor modalities
(pressure, inertial, and visual sensors). Various combinations and two sensor
fusion approaches (late and early fusion) for infant movement classification
were tested to evaluate whether a multi-sensor system outperforms single
modality assessments. Convolutional neural network (CNN) architectures were
used to classify movement patterns. The performance of the three-sensor fusion
(classification accuracy of 94.5%) was significantly higher than that of any
single modality evaluated. We show that the sensor fusion approach is a
promising avenue for automated classification of infant motor patterns. The
development of a robust sensor fusion system may significantly enhance AI-based
early recognition of neurofunctions, ultimately facilitating automated early
detection of neurodevelopmental conditions.
| [
{
"version": "v1",
"created": "Thu, 13 Jun 2024 11:38:58 GMT"
},
{
"version": "v2",
"created": "Fri, 14 Jun 2024 12:24:54 GMT"
},
{
"version": "v3",
"created": "Fri, 21 Jun 2024 05:24:28 GMT"
},
{
"version": "v4",
"created": "Tue, 23 Jul 2024 17:56:10 GMT"
},
{
"version": "v5",
"created": "Thu, 7 Nov 2024 15:41:04 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Kulvicius",
"Tomas",
""
],
[
"Zhang",
"Dajie",
""
],
[
"Poustka",
"Luise",
""
],
[
"Bölte",
"Sven",
""
],
[
"Jahn",
"Lennart",
""
],
[
"Flügge",
"Sarah",
""
],
[
"Kraft",
"Marc",
""
],
[
"Zweckstetter",
"Markus",
""
],
[
"Nielsen-Saines",
"Karin",
""
],
[
"Wörgötter",
"Florentin",
""
],
[
"Marschik",
"Peter B",
""
]
] |
2406.09215 | Yuxin Chen | Yuxin Chen, Junfei Tan, An Zhang, Zhengyi Yang, Leheng Sheng, Enzhi
Zhang, Xiang Wang, Tat-Seng Chua | On Softmax Direct Preference Optimization for Recommendation | NeurIPS 2024 | null | null | null | cs.IR cs.AI | http://creativecommons.org/licenses/by/4.0/ | Recommender systems aim to predict personalized rankings based on user
preference data. With the rise of Language Models (LMs), LM-based recommenders
have been widely explored due to their extensive world knowledge and powerful
reasoning abilities. Most of the LM-based recommenders convert historical
interactions into language prompts, pairing with a positive item as the target
response and fine-tuning LM with a language modeling loss. However, the current
objective fails to fully leverage preference data and is not optimized for
personalized ranking tasks, which hinders the performance of LM-based
recommenders. Inspired by the current advancement of Direct Preference
Optimization (DPO) in human preference alignment and the success of softmax
loss in recommendations, we propose Softmax-DPO (S-DPO) to instill ranking
information into the LM to help LM-based recommenders distinguish preferred
items from negatives, rather than solely focusing on positives. Specifically,
we incorporate multiple negatives in user preference data and devise an
alternative version of DPO loss tailored for LM-based recommenders, which is
extended from the traditional full-ranking Plackett-Luce (PL) model to partial
rankings and connected to softmax sampling strategies. Theoretically, we bridge
S-DPO with the softmax loss over negative sampling and find that it has an
inherent benefit of mining hard negatives, which assures its exceptional
capabilities in recommendation tasks. Empirically, extensive experiments
conducted on three real-world datasets demonstrate the superiority of S-DPO to
effectively model user preference and further boost recommendation performance
while providing better rewards for preferred items. Our codes are available at
https://github.com/chenyuxin1999/S-DPO.
| [
{
"version": "v1",
"created": "Thu, 13 Jun 2024 15:16:11 GMT"
},
{
"version": "v2",
"created": "Fri, 14 Jun 2024 15:22:58 GMT"
},
{
"version": "v3",
"created": "Thu, 7 Nov 2024 18:30:53 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Chen",
"Yuxin",
""
],
[
"Tan",
"Junfei",
""
],
[
"Zhang",
"An",
""
],
[
"Yang",
"Zhengyi",
""
],
[
"Sheng",
"Leheng",
""
],
[
"Zhang",
"Enzhi",
""
],
[
"Wang",
"Xiang",
""
],
[
"Chua",
"Tat-Seng",
""
]
] |
2406.09631 | Yuwei Wu | Yuwei Wu, Igor Spasojevic, Pratik Chaudhari, Vijay Kumar | Optimal Convex Cover as Collision-free Space Approximation for
Trajectory Generation | null | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose an online iterative algorithm to find an optimal convex cover to
under-approximate the free space for autonomous navigation to delineate Safe
Flight Corridors (SFC). The convex cover consists of a set of polytopes such
that the union of the polytopes represents obstacle-free space, allowing us to
find trajectories for robots that lie within the convex cover. In order to find
the SFC that facilitates optimal trajectory generation, we iteratively find
overlapping polytopes of maximum volumes that include specified waypoints
initialized by a geometric or kinematic planner. Constraints at waypoints
appear in two alternating stages of a joint optimization problem, which is
solved by a method inspired by the Alternating Direction Method of Multipliers
(ADMM) with partially distributed variables. We validate the effectiveness of
our proposed algorithm using a range of parameterized environments and show its
applications for two-stage motion planning.
| [
{
"version": "v1",
"created": "Thu, 13 Jun 2024 23:42:34 GMT"
},
{
"version": "v2",
"created": "Wed, 6 Nov 2024 20:08:48 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Wu",
"Yuwei",
""
],
[
"Spasojevic",
"Igor",
""
],
[
"Chaudhari",
"Pratik",
""
],
[
"Kumar",
"Vijay",
""
]
] |
2406.10129 | Mira Varma | Mira Varma and Oliver Baker | Investigating the Role of Quantum Entanglement in Heavy Ion Collisions
through Elliptic Flow | The presence of a thermal component doesn't necessarily imply
entanglement | null | null | null | nucl-th hep-ph quant-ph | http://creativecommons.org/licenses/by/4.0/ | This paper investigates the relationship between initial spatial anisotropy
and final state momentum anisotropy in heavy ion collisions through the
analysis of elliptic flow ($\nu_2$) as a function of transverse momentum
($p_T$). Building upon previous studies on thermalization in heavy ion
collisions using transverse momentum distributions, we extend the analysis to
the $p_T$ dependence of $\nu_2$ in Pb-Pb and Xe-Xe collisions. By employing a
two-component model to extract the thermal and hard scattering contributions to
the elliptic flow, we aim to gain further insights into the role of quantum
entanglement in the rapid thermalization and collective behavior of the
quark-gluon plasma (QGP).
| [
{
"version": "v1",
"created": "Fri, 14 Jun 2024 15:38:25 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 16:23:24 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Varma",
"Mira",
""
],
[
"Baker",
"Oliver",
""
]
] |
2406.10395 | Joseph Cox | Joseph Cox, Peng Liu, Skylar E. Stolte, Yunchao Yang, Kang Liu, Kyle
B. See, Huiwen Ju, Ruogu Fang | BrainSegFounder: Towards 3D Foundation Models for Neuroimage
Segmentation | 19 pages, 5 figures, to be published in Medical Image Analysis | null | null | null | eess.IV cs.CV q-bio.NC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The burgeoning field of brain health research increasingly leverages
artificial intelligence (AI) to interpret and analyze neurological data. This
study introduces a novel approach towards the creation of medical foundation
models by integrating a large-scale multi-modal magnetic resonance imaging
(MRI) dataset derived from 41,400 participants in its own. Our method involves
a novel two-stage pretraining approach using vision transformers. The first
stage is dedicated to encoding anatomical structures in generally healthy
brains, identifying key features such as shapes and sizes of different brain
regions. The second stage concentrates on spatial information, encompassing
aspects like location and the relative positioning of brain structures. We
rigorously evaluate our model, BrainFounder, using the Brain Tumor Segmentation
(BraTS) challenge and Anatomical Tracings of Lesions After Stroke v2.0 (ATLAS
v2.0) datasets. BrainFounder demonstrates a significant performance gain,
surpassing the achievements of the previous winning solutions using fully
supervised learning. Our findings underscore the impact of scaling up both the
complexity of the model and the volume of unlabeled training data derived from
generally healthy brains, which enhances the accuracy and predictive
capabilities of the model in complex neuroimaging tasks with MRI. The
implications of this research provide transformative insights and practical
applications in healthcare and make substantial steps towards the creation of
foundation models for Medical AI. Our pretrained models and training code can
be found at https://github.com/lab-smile/GatorBrain.
| [
{
"version": "v1",
"created": "Fri, 14 Jun 2024 19:49:45 GMT"
},
{
"version": "v2",
"created": "Mon, 12 Aug 2024 21:36:10 GMT"
},
{
"version": "v3",
"created": "Thu, 7 Nov 2024 15:28:50 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Cox",
"Joseph",
""
],
[
"Liu",
"Peng",
""
],
[
"Stolte",
"Skylar E.",
""
],
[
"Yang",
"Yunchao",
""
],
[
"Liu",
"Kang",
""
],
[
"See",
"Kyle B.",
""
],
[
"Ju",
"Huiwen",
""
],
[
"Fang",
"Ruogu",
""
]
] |
2406.10879 | Suchetan Das | Suchetan Das | Stretched Horizon from Conformal Field Theory | Published version in JHEP | null | null | null | hep-th | http://creativecommons.org/licenses/by/4.0/ | Recently, it has been observed that the Hartle-Hawking correlators, a
signature of smooth horizon, can emerge from certain heavy excited state
correlators in the (manifestly non-smooth) BTZ stretched horizon background, in
the limit when the stretched horizon approaches the real horizon. In this note,
we develop a framework of quantizing the CFT modular Hamiltonian, that explains
the necessity of introducing a stretched horizon and the emergence of thermal
features in the AdS-Rindler and (planar) BTZ backgrounds. In more detail, we
quantize vacuum modular Hamiltonian on a spatial segment of $S^{1}$. Unlike
radial quantization, (Euclidean) time circles emerge naturally here which can
be contracted smoothly to the `fixed points'(end points of the interval) of
this quantization thus providing a direct link to thermal physics. To define a
Hilbert space with discrete normalizable states and to construct a Virasoro
algebra with finite central extension, a natural regulator ($\epsilon$) is
needed around the fixed points. Eventually, in the dual description the fixed
points correspond to the horizons of AdS-Rindler patch or (planar) BTZ and the
cut-off being the stretched horizon. We construct a (Lorentzian) highest weight
representation of that Virasoro algebra. We further demonstrate that two point
function in a (vacuum) descendant state of the regulated Hilbert space will
reproduce thermal answer in $\epsilon \rightarrow 0$ limit which is analogous
to the recent observation of emergent thermality in stretched horizon
background. We also argue the thermal entropy of this quantization coincides
with entanglement entropy of the subregion. Conversely, the microcanonical
entropy corresponding to high energy density of states exactly reproduce the
BTZ entropy. Quite remarkably, all these dominant high lying microstates are
defined only at finite $\epsilon$ in the regulated Hilbert space.
| [
{
"version": "v1",
"created": "Sun, 16 Jun 2024 10:02:54 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 05:27:06 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Das",
"Suchetan",
""
]
] |
2406.11136 | Michael F. Xu | Michael F. Xu, Bengisu Cagiltay, Joseph Michaelis, Sarah Sebo, Bilge
Mutlu | Robots in Family Routines: Development of and Initial Insights from the
Family-Robot Routines Inventory | null | null | 10.1109/RO-MAN60168.2024.10731432 | null | cs.RO cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite advances in areas such as the personalization of robots, sustaining
adoption of robots for long-term use in families remains a challenge. Recent
studies have identified integrating robots into families' routines and rituals
as a promising approach to support long-term adoption. However, few studies
explored the integration of robots into family routines and there is a gap in
systematic measures to capture family preferences for robot integration.
Building upon existing routine inventories, we developed Family-Robot Routines
Inventory (FRRI), with 24 family routines and 24 child routine items, to
capture parents' attitudes toward and expectations from the integration of
robotic technology into their family routines. Using this inventory, we
collected data from 150 parents through an online survey. Our analysis
indicates that parents had varying perceptions for the utility of integrating
robots into their routines. For example, parents found robot integration to be
more helpful in children's individual routines, than to the collective routines
of their families. We discuss the design implications of these preliminary
findings, and how they may serve as a first step toward understanding the
diverse challenges and demands of designing and integrating household robots
for families.
| [
{
"version": "v1",
"created": "Mon, 17 Jun 2024 01:47:11 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Xu",
"Michael F.",
""
],
[
"Cagiltay",
"Bengisu",
""
],
[
"Michaelis",
"Joseph",
""
],
[
"Sebo",
"Sarah",
""
],
[
"Mutlu",
"Bilge",
""
]
] |
2406.11246 | Tomoki Matsumoto | Tomoki Matsumoto | Parallelizing MCMC with Machine Learning Classifier and Its Criterion
Based on Kullback-Leibler Divergence | null | null | null | null | stat.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the era of Big Data, Markov chain Monte Carlo (MCMC) methods, which are
currently essential for Bayesian estimation, face significant computational
challenges owing to their sequential nature. To achieve a faster and more
effective parallel computation, we emphasize the critical role of the
overlapped area of the posterior distributions based on partitioned data, which
we term the reconstructable area. We propose a method that utilizes machine
learning classifiers to effectively identify and extract MCMC draws obtained by
parallel computations from the area based on posteriors based on partitioned
sub-datasets, approximating the target posterior distribution based on the full
dataset. This study also develops a Kullback-Leibler (KL) divergence-based
criterion. It does not require calculating the full-posterior density and can
be calculated using only information from the sub-posterior densities, which
are generally obtained after implementing MCMC. This simplifies the
hyperparameter tuning in training classifiers. The simulation studies validated
the efficacy of the proposed method. This approach contributes to ongoing
research on parallelizing MCMC methods and may offer insights for future
developments in Bayesian computation for large-scale data analyses.
| [
{
"version": "v1",
"created": "Mon, 17 Jun 2024 06:17:31 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 05:42:52 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Matsumoto",
"Tomoki",
""
]
] |
2406.11709 | Priyanka Kargupta | Priyanka Kargupta, Ishika Agarwal, Dilek Hakkani-Tur, Jiawei Han | Instruct, Not Assist: LLM-based Multi-Turn Planning and Hierarchical
Questioning for Socratic Code Debugging | Code available at: https://github.com/agarwalishika/TreeInstruct
Accepted at EMNLP'24 Findings | null | null | null | cs.CL cs.MA | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Socratic questioning is an effective teaching strategy, encouraging critical
thinking and problem-solving. The conversational capabilities of large language
models (LLMs) show great potential for providing scalable, real-time student
guidance. However, current LLMs often give away solutions directly, making them
ineffective instructors. We tackle this issue in the code debugging domain with
TreeInstruct, an Instructor agent guided by a novel state space-based planning
algorithm. TreeInstruct asks probing questions to help students independently
identify and resolve errors. It estimates a student's conceptual and
syntactical knowledge to dynamically construct a question tree based on their
responses and current knowledge state, effectively addressing both independent
and dependent mistakes concurrently in a multi-turn interaction setting. In
addition to using an existing single-bug debugging benchmark, we construct a
more challenging multi-bug dataset of 150 coding problems, incorrect solutions,
and bug fixes -- all carefully constructed and annotated by experts. Extensive
evaluation shows TreeInstruct's state-of-the-art performance on both datasets,
proving it to be a more effective instructor than baselines. Furthermore, a
real-world case study with five students of varying skill levels further
demonstrates TreeInstruct's ability to guide students to debug their code
efficiently with minimal turns and highly Socratic questioning.
| [
{
"version": "v1",
"created": "Mon, 17 Jun 2024 16:28:21 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Jun 2024 21:17:41 GMT"
},
{
"version": "v3",
"created": "Tue, 20 Aug 2024 04:33:19 GMT"
},
{
"version": "v4",
"created": "Thu, 7 Nov 2024 07:00:14 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Kargupta",
"Priyanka",
""
],
[
"Agarwal",
"Ishika",
""
],
[
"Hakkani-Tur",
"Dilek",
""
],
[
"Han",
"Jiawei",
""
]
] |
2406.11779 | Jason Gross | Jason Gross, Rajashree Agrawal, Thomas Kwa, Euan Ong, Chun Hei Yip,
Alex Gibson, Soufiane Noubir, Lawrence Chan | Compact Proofs of Model Performance via Mechanistic Interpretability | accepted to the 38th Conference on Neural Information Processing
Systems (NeurIPS 2024) | null | null | null | cs.LG cs.LO | http://creativecommons.org/licenses/by/4.0/ | We propose using mechanistic interpretability -- techniques for reverse
engineering model weights into human-interpretable algorithms -- to derive and
compactly prove formal guarantees on model performance. We prototype this
approach by formally proving accuracy lower bounds for a small transformer
trained on Max-of-K, validating proof transferability across 151 random seeds
and four values of K. We create 102 different computer-assisted proof
strategies and assess their length and tightness of bound on each of our
models. Using quantitative metrics, we find that shorter proofs seem to require
and provide more mechanistic understanding. Moreover, we find that more
faithful mechanistic understanding leads to tighter performance bounds. We
confirm these connections by qualitatively examining a subset of our proofs.
Finally, we identify compounding structureless errors as a key challenge for
using mechanistic interpretability to generate compact proofs on model
performance.
| [
{
"version": "v1",
"created": "Mon, 17 Jun 2024 17:34:25 GMT"
},
{
"version": "v10",
"created": "Fri, 6 Sep 2024 02:08:07 GMT"
},
{
"version": "v11",
"created": "Thu, 31 Oct 2024 08:24:40 GMT"
},
{
"version": "v12",
"created": "Thu, 7 Nov 2024 00:11:55 GMT"
},
{
"version": "v2",
"created": "Tue, 18 Jun 2024 12:36:07 GMT"
},
{
"version": "v3",
"created": "Wed, 19 Jun 2024 07:52:58 GMT"
},
{
"version": "v4",
"created": "Fri, 21 Jun 2024 17:58:28 GMT"
},
{
"version": "v5",
"created": "Mon, 24 Jun 2024 13:06:01 GMT"
},
{
"version": "v6",
"created": "Sun, 30 Jun 2024 02:00:46 GMT"
},
{
"version": "v7",
"created": "Sun, 7 Jul 2024 19:19:51 GMT"
},
{
"version": "v8",
"created": "Fri, 12 Jul 2024 21:51:34 GMT"
},
{
"version": "v9",
"created": "Sun, 21 Jul 2024 18:30:20 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Gross",
"Jason",
""
],
[
"Agrawal",
"Rajashree",
""
],
[
"Kwa",
"Thomas",
""
],
[
"Ong",
"Euan",
""
],
[
"Yip",
"Chun Hei",
""
],
[
"Gibson",
"Alex",
""
],
[
"Noubir",
"Soufiane",
""
],
[
"Chan",
"Lawrence",
""
]
] |
2406.11944 | Jacob Dunefsky | Jacob Dunefsky and Philippe Chlenski and Neel Nanda | Transcoders Find Interpretable LLM Feature Circuits | 29 pages, 6 figures, 4 tables, 2 algorithms. NeurIPS 2024 | null | null | null | cs.LG cs.CL | http://creativecommons.org/licenses/by/4.0/ | A key goal in mechanistic interpretability is circuit analysis: finding
sparse subgraphs of models corresponding to specific behaviors or capabilities.
However, MLP sublayers make fine-grained circuit analysis on transformer-based
language models difficult. In particular, interpretable features -- such as
those found by sparse autoencoders (SAEs) -- are typically linear combinations
of extremely many neurons, each with its own nonlinearity to account for.
Circuit analysis in this setting thus either yields intractably large circuits
or fails to disentangle local and global behavior. To address this we explore
transcoders, which seek to faithfully approximate a densely activating MLP
layer with a wider, sparsely-activating MLP layer. We introduce a novel method
for using transcoders to perform weights-based circuit analysis through MLP
sublayers. The resulting circuits neatly factorize into input-dependent and
input-invariant terms. We then successfully train transcoders on language
models with 120M, 410M, and 1.4B parameters, and find them to perform at least
on par with SAEs in terms of sparsity, faithfulness, and
human-interpretability. Finally, we apply transcoders to reverse-engineer
unknown circuits in the model, and we obtain novel insights regarding the
"greater-than circuit" in GPT2-small. Our results suggest that transcoders can
prove effective in decomposing model computations involving MLPs into
interpretable circuits. Code is available at
https://github.com/jacobdunefsky/transcoder_circuits/.
| [
{
"version": "v1",
"created": "Mon, 17 Jun 2024 17:49:00 GMT"
},
{
"version": "v2",
"created": "Wed, 6 Nov 2024 22:37:30 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Dunefsky",
"Jacob",
""
],
[
"Chlenski",
"Philippe",
""
],
[
"Nanda",
"Neel",
""
]
] |
2406.12895 | Marco Cafiso | Marco Cafiso and Paolo Paradisi | Temporal Complexity of a Hopfield-Type Neural Model in Random and
Scale-Free Graphs | null | null | null | null | q-bio.NC cond-mat.dis-nn cs.NA math-ph math.MP math.NA nlin.AO | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The Hopfield network model and its generalizations were introduced as a model
of associative, or content-addressable, memory. They were widely investigated
both as an unsupervised learning method in artificial intelligence and as a
model of biological neural dynamics in computational neuroscience. The
complexity features of biological neural networks have attracted the scientific
community's interest for the last two decades. More recently, concepts and
tools borrowed from complex network theory were applied to artificial neural
networks and learning, thus focusing on the topological aspects. However, the
temporal structure is also a crucial property displayed by biological neural
networks and investigated in the framework of systems displaying complex
intermittency. The Intermittency-Driven Complexity (IDC) approach indeed
focuses on the metastability of self-organized states, whose signature is a
power-decay in the inter-event time distribution or a scaling behaviour in the
related event-driven diffusion processes. The investigation of IDC in neural
dynamics and its relationship with network topology is still in its early
stages. In this work, we present the preliminary results of an IDC analysis
carried out on a bio-inspired Hopfield-type neural network comparing two
different connectivities, i.e., scale-free vs. random network topology. We
found that random networks can trigger complexity features similar to that of
scale-free networks, even if with some differences and for different parameter
values, in particular for different noise levels
| [
{
"version": "v1",
"created": "Thu, 6 Jun 2024 17:57:38 GMT"
},
{
"version": "v2",
"created": "Fri, 19 Jul 2024 15:34:34 GMT"
},
{
"version": "v3",
"created": "Thu, 7 Nov 2024 11:48:15 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Cafiso",
"Marco",
""
],
[
"Paradisi",
"Paolo",
""
]
] |
2406.14894 | Andrea Tagarelli | Candida M. Greco, Lucio La Cava, Andrea Tagarelli | Talking the Talk Does Not Entail Walking the Walk: On the Limits of
Large Language Models in Lexical Entailment Recognition | Accepted for publication at The 2024 Conference on Empirical Methods
in Natural Language Processing (EMNLP-2024) - Findings | null | null | null | cs.CL cs.AI cs.CY cs.IR physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Verbs form the backbone of language, providing the structure and meaning to
sentences. Yet, their intricate semantic nuances pose a longstanding challenge.
Understanding verb relations through the concept of lexical entailment is
crucial for comprehending sentence meanings and grasping verb dynamics. This
work investigates the capabilities of eight Large Language Models in
recognizing lexical entailment relations among verbs through differently
devised prompting strategies and zero-/few-shot settings over verb pairs from
two lexical databases, namely WordNet and HyperLex. Our findings unveil that
the models can tackle the lexical entailment recognition task with moderately
good performance, although at varying degree of effectiveness and under
different conditions. Also, utilizing few-shot prompting can enhance the
models' performance. However, perfectly solving the task arises as an unmet
challenge for all examined LLMs, which raises an emergence for further research
developments on this topic.
| [
{
"version": "v1",
"created": "Fri, 21 Jun 2024 06:30:16 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 18:15:23 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Greco",
"Candida M.",
""
],
[
"La Cava",
"Lucio",
""
],
[
"Tagarelli",
"Andrea",
""
]
] |
2406.14920 | Diptarka Hait | Yuanheng Wang and Diptarka Hait and K. Grace Johnson and O. Jonathan
Fajen and Juncheng Harry Zhang and Rub\'en D. Guerrero and Todd J. Mart\'inez | Extending GPU-Accelerated Gaussian Integrals in the TeraChem Software
Package to f Type Orbitals: Implementation and Applications | Added additional data | J. Chem. Phys. 161, 174118 (2024) | 10.1063/5.0233523 | null | physics.chem-ph physics.comp-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The increasing availability of GPUs for scientific computing has prompted
interest in accelerating quantum chemical calculations through their use. The
complexity of integral kernels for high angular momentum basis functions
however often limits the utility of GPU implementations with large basis sets
or for metal containing systems. In this work, we report implementation of $f$
function support in the GPU-accelerated TeraChem software package through the
development of efficient kernels for the evaluation of Hamiltonian integrals.
The high efficiency of the resulting code is demonstrated through density
functional theory (DFT) calculations on increasingly large organic molecules
and transition metal complexes, as well as coupled cluster singles and doubles
(CCSD) calculations on water clusters. Preliminary investigations into Ni(I)
catalysis with DFT and the photochemistry of MnH(CH$_3$) with complete active
space self-consistent field (CASSCF) are also carried out. Overall, our
GPU-accelerated software appears to be well-suited for fast simulation of large
transition metal containing systems, as well as organic molecules.
| [
{
"version": "v1",
"created": "Fri, 21 Jun 2024 07:29:50 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Aug 2024 19:28:12 GMT"
},
{
"version": "v3",
"created": "Sun, 13 Oct 2024 19:24:42 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Wang",
"Yuanheng",
""
],
[
"Hait",
"Diptarka",
""
],
[
"Johnson",
"K. Grace",
""
],
[
"Fajen",
"O. Jonathan",
""
],
[
"Zhang",
"Juncheng Harry",
""
],
[
"Guerrero",
"Rubén D.",
""
],
[
"Martínez",
"Todd J.",
""
]
] |
2406.15586 | Zachary Horvitz | Zachary Horvitz, Ajay Patel, Kanishk Singh, Chris Callison-Burch,
Kathleen McKeown, Zhou Yu | TinyStyler: Efficient Few-Shot Text Style Transfer with Authorship
Embeddings | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | The goal of text style transfer is to transform the style of texts while
preserving their original meaning, often with only a few examples of the target
style. Existing style transfer methods generally rely on the few-shot
capabilities of large language models or on complex controllable text
generation approaches that are inefficient and underperform on fluency metrics.
We introduce TinyStyler, a lightweight but effective approach, which leverages
a small language model (800M params) and pre-trained authorship embeddings to
perform efficient, few-shot text style transfer. We evaluate on the challenging
task of authorship style transfer and find TinyStyler outperforms strong
approaches such as GPT-4. We also evaluate TinyStyler's ability to perform text
attribute style transfer (formal $\leftrightarrow$ informal) with automatic and
human evaluations and find that the approach outperforms recent controllable
text generation methods. Our model has been made publicly available at
https://huggingface.co/tinystyler/tinystyler .
| [
{
"version": "v1",
"created": "Fri, 21 Jun 2024 18:41:22 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 17:56:19 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Horvitz",
"Zachary",
""
],
[
"Patel",
"Ajay",
""
],
[
"Singh",
"Kanishk",
""
],
[
"Callison-Burch",
"Chris",
""
],
[
"McKeown",
"Kathleen",
""
],
[
"Yu",
"Zhou",
""
]
] |
2406.15708 | Xingchen Wan | Xingchen Wan, Ruoxi Sun, Hootan Nakhost, Sercan O. Arik | Teach Better or Show Smarter? On Instructions and Exemplars in Automatic
Prompt Optimization | Expanded version of the NeurIPS 2024 paper | null | null | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models have demonstrated remarkable capabilities, but their
performance is heavily reliant on effective prompt engineering. Automatic
prompt optimization (APO) methods are designed to automate this and can be
broadly categorized into those targeting instructions (instruction
optimization, IO) vs. those targeting exemplars (exemplar optimization, EO).
Despite their shared objective, these have evolved rather independently, with
IO receiving more research attention recently. This paper seeks to bridge this
gap by comprehensively comparing the performance of representative IO and EO
techniques both isolation and combination on a diverse set of challenging
tasks. Our findings reveal that intelligently reusing model-generated
input-output pairs obtained from evaluating prompts on the validation set as
exemplars, consistently improves performance on top of IO methods but is
currently under-investigated. We also find that despite the recent focus on IO,
how we select exemplars can outweigh how we optimize instructions, with EO
strategies as simple as random search outperforming state-of-the-art IO methods
with seed instructions without any optimization. Moreover, we observe a synergy
between EO and IO, with optimal combinations surpassing the individual
contributions. We conclude that studying exemplar optimization both as a
standalone method and its optimal combination with instruction optimization
remain a crucial aspect of APO and deserve greater consideration in future
research, even in the era of highly capable instruction-following models.
| [
{
"version": "v1",
"created": "Sat, 22 Jun 2024 02:07:10 GMT"
},
{
"version": "v2",
"created": "Wed, 6 Nov 2024 22:07:17 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Wan",
"Xingchen",
""
],
[
"Sun",
"Ruoxi",
""
],
[
"Nakhost",
"Hootan",
""
],
[
"Arik",
"Sercan O.",
""
]
] |
2406.15818 | Yermek Aldabergenov | Yermek Aldabergenov, Ding Ding, Wei Lin, Yidun Wan | Meso-inflationary Peccei-Quinn symmetry breaking with non-minimal
coupling | Updated version (to appear in ApJS). 34 pages, 11 figures, 1 table | null | 10.3847/1538-4365/ad77bd | null | hep-th astro-ph.CO gr-qc hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study a realization of the inflationary scenario where the Peccei-Quinn
(PQ) symmetry is spontaneously broken during inflation, facilitated by its
non-minimal coupling to gravity. This results in effectively two-field
inflation: the early stage is driven by an inflaton field with the PQ symmetry
intact, and the later stage is driven by the PQ scalar after its effective mass
becomes tachyonic, causing destabilization from the origin. The non-minimal
coupling serves the dual purpose of restoring the PQ symmetry during early
inflation and flattening the PQ potential post-tachyonic shift, allowing for
continued slow roll. We analyze the inflationary background solutions and
scalar perturbations, which are amplified at small scales via significant
isocurvature perturbations generated near the symmetry-breaking epoch. These
perturbations lead to second-order gravitational waves, detectable by
next-generation space-based experiments.
| [
{
"version": "v1",
"created": "Sat, 22 Jun 2024 11:03:08 GMT"
},
{
"version": "v2",
"created": "Thu, 12 Sep 2024 19:54:05 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Aldabergenov",
"Yermek",
""
],
[
"Ding",
"Ding",
""
],
[
"Lin",
"Wei",
""
],
[
"Wan",
"Yidun",
""
]
] |
2406.15830 | Haru-Tada Sato | Haru-Tada Sato (Osaka Metropolitan University & i's Factory Co., Ltd.) | Moyal product and Generalized Hom-Lie-Virasoro symmetries in Bloch
electron systems | Latex, 35 pages | Nucl, Phys. B1006 (2024) 116639 | 10.1016/j.nuclphysb.2024.116639 | null | hep-th math-ph math.MP | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We explore two variations of the Curtright-Zachos (CZ) deformation of the
Virasoro algebra. Firstly, we introduce a scaled CZ algebra that inherits the
scaling structure found in the differential operator representation of the
magnetic translation (MT) operators. We then linearly decompose the scaled CZ
generators to derive two types of Hom-Lie deformations of the W-infinity
algebra. We discuss *-bracket formulations of these algebras and their
connection to the Moyal product. We show that the *-bracket form of the scaled
CZ algebra arises from the Moyal product, while we obtain the second type of
deformed W-infinity through a coordinate transformation of the first type of
Moyal operators. From a physical point of view, we construct the Hamiltonian of
a tight binding model (TBM) using the Wyle matrix representation of the scaled
CZ algebra. We note that the integer powers of q are linked to the quantum
fluctuations that are inherent in the Moyal product.
| [
{
"version": "v1",
"created": "Sat, 22 Jun 2024 12:21:47 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Sato",
"Haru-Tada",
"",
"Osaka Metropolitan University & i's Factory Co., Ltd."
]
] |
2406.16148 | Yuwei Zhang | Yuwei Zhang, Tong Xia, Jing Han, Yu Wu, Georgios Rizos, Yang Liu,
Mohammed Mosuily, Jagmohan Chauhan, Cecilia Mascolo | Towards Open Respiratory Acoustic Foundation Models: Pretraining and
Benchmarking | accepted by NeurIPS 2024 Track Datasets and Benchmarks | null | null | null | cs.SD cs.AI cs.LG eess.AS | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Respiratory audio, such as coughing and breathing sounds, has predictive
power for a wide range of healthcare applications, yet is currently
under-explored. The main problem for those applications arises from the
difficulty in collecting large labeled task-specific data for model
development. Generalizable respiratory acoustic foundation models pretrained
with unlabeled data would offer appealing advantages and possibly unlock this
impasse. However, given the safety-critical nature of healthcare applications,
it is pivotal to also ensure openness and replicability for any proposed
foundation model solution. To this end, we introduce OPERA, an OPEn Respiratory
Acoustic foundation model pretraining and benchmarking system, as the first
approach answering this need. We curate large-scale respiratory audio datasets
(~136K samples, over 400 hours), pretrain three pioneering foundation models,
and build a benchmark consisting of 19 downstream respiratory health tasks for
evaluation. Our pretrained models demonstrate superior performance (against
existing acoustic models pretrained with general audio on 16 out of 19 tasks)
and generalizability (to unseen datasets and new respiratory audio modalities).
This highlights the great promise of respiratory acoustic foundation models and
encourages more studies using OPERA as an open resource to accelerate research
on respiratory audio for health. The system is accessible from
https://github.com/evelyn0414/OPERA.
| [
{
"version": "v1",
"created": "Sun, 23 Jun 2024 16:04:26 GMT"
},
{
"version": "v2",
"created": "Mon, 28 Oct 2024 15:05:04 GMT"
},
{
"version": "v3",
"created": "Thu, 7 Nov 2024 15:23:59 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Zhang",
"Yuwei",
""
],
[
"Xia",
"Tong",
""
],
[
"Han",
"Jing",
""
],
[
"Wu",
"Yu",
""
],
[
"Rizos",
"Georgios",
""
],
[
"Liu",
"Yang",
""
],
[
"Mosuily",
"Mohammed",
""
],
[
"Chauhan",
"Jagmohan",
""
],
[
"Mascolo",
"Cecilia",
""
]
] |
2406.16700 | Yang Qi | Zhaoyang Ding and Yang Qi | Anomalies in mirror symmetry enriched topological orders | 13 pages, 5 figures | Phys. Rev. B 110, 155151 (2024) | 10.1103/PhysRevB.110.155151 | null | cond-mat.str-el hep-th math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Two-dimensional mirror symmetry enriched topological (SET) orders can be
studied using the folding approach: it can be folded along the mirror axis and
turned into a bilayer system on which the mirror symmetry acts as a $\mathbb
Z_2$ layer-exchange symmetry. How mirror symmetry enriches the topological
order is then encoded at the mirror axis, which is a gapped boundary of the
folded bilayer system. Based on anyon-condensation theory, we classify possible
$\mathbb Z_2$-symmetric gapped boundaries of the folded system. In particular,
we derive an $H^2$ obstruction function, which corresponds to an $H^3$
obstruction for topological orders enriched by the time-reversal symmetry
instead of mirror symmetry. We demonstrate that states with a nontrivial $H^2$
obstruction function can be constructed on the surface of a three-dimensional
mirror SET order.
| [
{
"version": "v1",
"created": "Mon, 24 Jun 2024 15:02:21 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 13:48:05 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Ding",
"Zhaoyang",
""
],
[
"Qi",
"Yang",
""
]
] |
2406.17052 | Andrew Santarelli | Andrew D. Santarelli, Matthew E. Caplan, Earl P. Bellinger | Formation of Sub-Chandrasekhar Mass Black Holes and Red Stragglers via
Hawking Stars in Ultra-Faint Dwarf Galaxies | 12 pages, 5 figures, submitted to and accepted by the ApJ | null | null | null | astro-ph.GA astro-ph.SR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Presently, primordial black holes (PBHs) in the asteroid-mass window from
$10^{-16}$ M$_\odot$ to $10^{-10}$ M$_\odot$ are a popular dark matter
candidate. If they exist, some stars would capture them upon formation, and
they would slowly accrete the star over gigayears. Such Hawking stars -- stars
with a central PBH -- provide a novel channel for the formation of both
sub-Chandrasekhar mass black holes as well as red straggler stars. Here we
report on stellar evolution models that extend our previous work to Hawking
stars with masses between 0.5 and 1.4 M$_\odot$. We explore three accretion
schemes, and find that a wide range of PBHs in the asteroid-mass window can
robustly accrete stars as small as 1 M$_\odot$ within the age of the Universe.
This mechanism of producing sub-solar mass black holes is highly dependent on
the assumed accretion physics and stellar metallicity. Lower-metallicity stars
are generally accreted more rapidly, suggesting that it may be more likely for
sub-Chandrasekhar mass Hawking stars formed in the early universe, such as
those in ultra-faint dwarf (UFD) galaxies, to transmute their star into a
sub-Chandrasekhar mass black hole within a Hubble time. We present a stellar
population synthesis of a Draco II-like UFD galaxy containing Hawking stars and
show that the number of red stragglers they produce can qualitatively match the
observed population \mb{for black hole seed masses around 10$^{-11}$ M$_\odot$
and under the assumption that they accrete with high radiative efficiency}.
| [
{
"version": "v1",
"created": "Mon, 24 Jun 2024 18:14:31 GMT"
},
{
"version": "v2",
"created": "Wed, 6 Nov 2024 20:21:05 GMT"
}
] | 2024-11-08T00:00:00 | [
[
"Santarelli",
"Andrew D.",
""
],
[
"Caplan",
"Matthew E.",
""
],
[
"Bellinger",
"Earl P.",
""
]
] |