date
timestamp[ns]date 2023-05-05 00:00:00
2025-03-28 00:00:00
| arxiv_id
stringlengths 10
10
| title
stringlengths 8
177
| authors
sequencelengths 1
942
| github
stringlengths 0
116
| abstract
stringlengths 165
1.92k
|
|---|---|---|---|---|---|
2025-03-19T00:00:00 | 2503.12545 | PEBench: A Fictitious Dataset to Benchmark Machine Unlearning for Multimodal Large Language Models | [
"Zhaopan Xu",
"Pengfei Zhou",
"Weidong Tang",
"Jiaxin Ai",
"Wangbo Zhao",
"Xiaojiang Peng",
"Kai Wang",
"Yang You",
"Wenqi Shao",
"Hongxun Yao",
"Kaipeng Zhang"
] | In recent years, Multimodal Large Language Models (MLLMs) have demonstrated remarkable advancements in tasks such as visual question answering, visual understanding, and reasoning. However, this impressive progress relies on vast amounts of data collected from the internet, raising significant concerns about privacy and security. To address these issues, machine unlearning (MU) has emerged as a promising solution, enabling the removal of specific knowledge from an already trained model without requiring retraining from scratch. Although MU for MLLMs has gained attention, current evaluations of its efficacy remain incomplete, and the underlying problem is often poorly defined, which hinders the development of strategies for creating more secure and trustworthy systems. To bridge this gap, we introduce a benchmark, named PEBench, which includes a dataset of personal entities and corresponding general event scenes, designed to comprehensively assess the performance of MU for MLLMs. Through PEBench, we aim to provide a standardized and robust framework to advance research in secure and privacy-preserving multimodal models. We benchmarked 6 MU methods, revealing their strengths and limitations, and shedding light on key challenges and opportunities for MU in MLLMs. |
|
2025-03-19T00:00:00 | 2503.12505 | MPBench: A Comprehensive Multimodal Reasoning Benchmark for Process Errors Identification | [
"Zhaopan Xu",
"Pengfei Zhou",
"Jiaxin Ai",
"Wangbo Zhao",
"Kai Wang",
"Xiaojiang Peng",
"Wenqi Shao",
"Hongxun Yao",
"Kaipeng Zhang"
] | Reasoning is an essential capacity for large language models (LLMs) to address complex tasks, where the identification of process errors is vital for improving this ability. Recently, process-level reward models (PRMs) were proposed to provide step-wise rewards that facilitate reinforcement learning and data production during training and guide LLMs toward correct steps during inference, thereby improving reasoning accuracy. However, existing benchmarks of PRMs are text-based and focus on error detection, neglecting other scenarios like reasoning search. To address this gap, we introduce MPBench, a comprehensive, multi-task, multimodal benchmark designed to systematically assess the effectiveness of PRMs in diverse scenarios. MPBench employs three evaluation paradigms, each targeting a specific role of PRMs in the reasoning process: (1) Step Correctness, which assesses the correctness of each intermediate reasoning step; (2) Answer Aggregation, which aggregates multiple solutions and selects the best one; and (3) Reasoning Process Search, which guides the search for optimal reasoning steps during inference. Through these paradigms, MPBench makes comprehensive evaluations and provides insights into the development of multimodal PRMs. |
|
2025-03-19T00:00:00 | 2503.10410 | RoCo-Sim: Enhancing Roadside Collaborative Perception through Foreground Simulation | [
"Yuwen Du",
"Anning Hu",
"Zichen Chao",
"Yifan Lu",
"Junhao Ge",
"Genjia Liu",
"Weitao Wu",
"Lanjun Wang",
"Siheng Chen"
] | https://github.com/duyuwen-duen/RoCo-Sim | Roadside Collaborative Perception refers to a system where multiple roadside units collaborate to pool their perceptual data, assisting vehicles in enhancing their environmental awareness. Existing roadside perception methods concentrate on model design but overlook data issues like calibration errors, sparse information, and multi-view consistency, leading to poor performance on recent published datasets. To significantly enhance roadside collaborative perception and address critical data issues, we present the first simulation framework RoCo-Sim for road-side collaborative perception. RoCo-Sim is capable of generating diverse, multi-view consistent simulated roadside data through dynamic foreground editing and full-scene style transfer of a single image. RoCo-Sim consists of four components: (1) Camera Extrinsic Optimization ensures accurate 3D to 2D projection for roadside cameras; (2) A novel Multi-View Occlusion-Aware Sampler (MOAS) determines the placement of diverse digital assets within 3D space; (3) DepthSAM innovatively models foreground-background relationships from single-frame fixed-view images, ensuring multi-view consistency of foreground; and (4) Scalable Post-Processing Toolkit generates more realistic and enriched scenes through style transfer and other enhancements. RoCo-Sim significantly improves roadside 3D object detection, outperforming SOTA methods by 83.74 on Rcooper-Intersection and 83.12 on TUMTraf-V2X for AP70. RoCo-Sim fills a critical gap in roadside perception simulation. Code and pre-trained models will be released soon: https://github.com/duyuwen-duen/RoCo-Sim |
2025-03-19T00:00:00 | 2503.10546 | KUDA: Keypoints to Unify Dynamics Learning and Visual Prompting for Open-Vocabulary Robotic Manipulation | [
"Zixian Liu",
"Mingtong Zhang",
"Yunzhu Li"
] | With the rapid advancement of large language models (LLMs) and vision-language models (VLMs), significant progress has been made in developing open-vocabulary robotic manipulation systems. However, many existing approaches overlook the importance of object dynamics, limiting their applicability to more complex, dynamic tasks. In this work, we introduce KUDA, an open-vocabulary manipulation system that integrates dynamics learning and visual prompting through keypoints, leveraging both VLMs and learning-based neural dynamics models. Our key insight is that a keypoint-based target specification is simultaneously interpretable by VLMs and can be efficiently translated into cost functions for model-based planning. Given language instructions and visual observations, KUDA first assigns keypoints to the RGB image and queries the VLM to generate target specifications. These abstract keypoint-based representations are then converted into cost functions, which are optimized using a learned dynamics model to produce robotic trajectories. We evaluate KUDA on a range of manipulation tasks, including free-form language instructions across diverse object categories, multi-object interactions, and deformable or granular objects, demonstrating the effectiveness of our framework. The project page is available at http://kuda-dynamics.github.io. |
|
2025-03-19T00:00:00 | 2503.14456 | RWKV-7 "Goose" with Expressive Dynamic State Evolution | [
"Bo Peng",
"Ruichong Zhang",
"Daniel Goldstein",
"Eric Alcaide",
"Haowen Hou",
"Janna Lu",
"William Merrill",
"Guangyu Song",
"Kaifeng Tan",
"Saiteja Utpala",
"Nathan Wilce",
"Johan S. Wind",
"Tianyi Wu",
"Daniel Wuttke",
"Christian Zhou-Zheng"
] | https://github.com/RWKV/RWKV-LM | We present RWKV-7 "Goose", a new sequence modeling architecture, along with pre-trained language models that establish a new state-of-the-art in downstream performance at the 3 billion parameter scale on multilingual tasks, and match current SoTA English language performance despite being trained on dramatically fewer tokens than other top 3B models. Nevertheless, RWKV-7 models require only constant memory usage and constant inference time per token. RWKV-7 introduces a newly generalized formulation of the delta rule with vector-valued gating and in-context learning rates, as well as a relaxed value replacement rule. We show that RWKV-7 can perform state tracking and recognize all regular languages, while retaining parallelizability of training. This exceeds the capabilities of Transformers under standard complexity conjectures, which are limited to TC^0. To demonstrate RWKV-7's language modeling capability, we also present an extended open source 3.1 trillion token multilingual corpus, and train four RWKV-7 models ranging from 0.19 billion to 2.9 billion parameters on this dataset. To foster openness, reproduction, and adoption, we release our models and dataset component listing at https://huggingface.co/RWKV, and our training and inference code at https://github.com/RWKV/RWKV-LM all under the Apache 2.0 License. |
2025-03-19T00:00:00 | 2503.13424 | Infinite Mobility: Scalable High-Fidelity Synthesis of Articulated Objects via Procedural Generation | [
"Xinyu Lian",
"Zichao Yu",
"Ruiming Liang",
"Yitong Wang",
"Li Ray Luo",
"Kaixu Chen",
"Yuanzhen Zhou",
"Qihong Tang",
"Xudong Xu",
"Zhaoyang Lyu",
"Bo Dai",
"Jiangmiao Pang"
] | https://github.com/Intern-Nexus/Infinite-Mobility | Large-scale articulated objects with high quality are desperately needed for multiple tasks related to embodied AI. Most existing methods for creating articulated objects are either data-driven or simulation based, which are limited by the scale and quality of the training data or the fidelity and heavy labour of the simulation. In this paper, we propose Infinite Mobility, a novel method for synthesizing high-fidelity articulated objects through procedural generation. User study and quantitative evaluation demonstrate that our method can produce results that excel current state-of-the-art methods and are comparable to human-annotated datasets in both physics property and mesh quality. Furthermore, we show that our synthetic data can be used as training data for generative models, enabling next-step scaling up. Code is available at https://github.com/Intern-Nexus/Infinite-Mobility |
2025-03-19T00:00:00 | 2503.12797 | DeepPerception: Advancing R1-like Cognitive Visual Perception in MLLMs for Knowledge-Intensive Visual Grounding | [
"Xinyu Ma",
"Ziyang Ding",
"Zhicong Luo",
"Chi Chen",
"Zonghao Guo",
"Derek F. Wong",
"Xiaoyi Feng",
"Maosong Sun"
] | https://github.com/thunlp/DeepPerception | Human experts excel at fine-grained visual discrimination by leveraging domain knowledge to refine perceptual features, a capability that remains underdeveloped in current Multimodal Large Language Models (MLLMs). Despite possessing vast expert-level knowledge, MLLMs struggle to integrate reasoning into visual perception, often generating direct responses without deeper analysis. To bridge this gap, we introduce knowledge-intensive visual grounding (KVG), a novel visual grounding task that requires both fine-grained perception and domain-specific knowledge integration. To address the challenges of KVG, we propose DeepPerception, an MLLM enhanced with cognitive visual perception capabilities. Our approach consists of (1) an automated data synthesis pipeline that generates high-quality, knowledge-aligned training samples, and (2) a two-stage training framework combining supervised fine-tuning for cognitive reasoning scaffolding and reinforcement learning to optimize perception-cognition synergy. To benchmark performance, we introduce KVG-Bench a comprehensive dataset spanning 10 domains with 1.3K manually curated test cases. Experimental results demonstrate that DeepPerception significantly outperforms direct fine-tuning, achieving +8.08\% accuracy improvements on KVG-Bench and exhibiting +4.60\% superior cross-domain generalization over baseline approaches. Our findings highlight the importance of integrating cognitive processes into MLLMs for human-like visual perception and open new directions for multimodal reasoning research. The data, codes, and models are released at https://github.com/thunlp/DeepPerception. |
2025-03-19T00:00:00 | 2503.14476 | DAPO: An Open-Source LLM Reinforcement Learning System at Scale | [
"Qiying Yu",
"Zheng Zhang",
"Ruofei Zhu",
"Yufeng Yuan",
"Xiaochen Zuo",
"Yu Yue",
"Tiantian Fan",
"Gaohong Liu",
"Lingjun Liu",
"Xin Liu",
"Haibin Lin",
"Zhiqi Lin",
"Bole Ma",
"Guangming Sheng",
"Yuxuan Tong",
"Chi Zhang",
"Mofan Zhang",
"Wang Zhang",
"Hang Zhu",
"Jinhua Zhu",
"Jiaze Chen",
"Jiangjie Chen",
"Chengyi Wang",
"Hongli Yu",
"Weinan Dai",
"Yuxuan Song",
"Xiangpeng Wei",
"Hao Zhou",
"Jingjing Liu",
"Wei-Ying Ma",
"Ya-Qin Zhang",
"Lin Yan",
"Mu Qiao",
"Yonghui Wu",
"Mingxuan Wang"
] | Inference scaling empowers LLMs with unprecedented reasoning ability, with reinforcement learning as the core technique to elicit complex reasoning. However, key technical details of state-of-the-art reasoning LLMs are concealed (such as in OpenAI o1 blog and DeepSeek R1 technical report), thus the community still struggles to reproduce their RL training results. We propose the Decoupled Clip and Dynamic sAmpling Policy Optimization (DAPO) algorithm, and fully open-source a state-of-the-art large-scale RL system that achieves 50 points on AIME 2024 using Qwen2.5-32B base model. Unlike previous works that withhold training details, we introduce four key techniques of our algorithm that make large-scale LLM RL a success. In addition, we open-source our training code, which is built on the verl framework, along with a carefully curated and processed dataset. These components of our open-source system enhance reproducibility and support future research in large-scale LLM RL. |
|
2025-03-19T00:00:00 | 2503.14499 | Measuring AI Ability to Complete Long Tasks | [
"Thomas Kwa",
"Ben West",
"Joel Becker",
"Amy Deng",
"Katharyn Garcia",
"Max Hasin",
"Sami Jawhar",
"Megan Kinniment",
"Nate Rush",
"Sydney Von Arx",
"Ryan Bloom",
"Thomas Broadley",
"Haoxing Du",
"Brian Goodrich",
"Nikola Jurkovic",
"Luke Harold Miles",
"Seraphina Nix",
"Tao Lin",
"Neev Parikh",
"David Rein",
"Lucas Jun Koba Sato",
"Hjalmar Wijk",
"Daniel M. Ziegler",
"Elizabeth Barnes",
"Lawrence Chan"
] | Despite rapid progress on AI benchmarks, the real-world meaning of benchmark performance remains unclear. To quantify the capabilities of AI systems in terms of human capabilities, we propose a new metric: 50%-task-completion time horizon. This is the time humans typically take to complete tasks that AI models can complete with 50% success rate. We first timed humans with relevant domain expertise on a combination of RE-Bench, HCAST, and 66 novel shorter tasks. On these tasks, current frontier AI models such as Claude 3.7 Sonnet have a 50% time horizon of around 50 minutes. Furthermore, frontier AI time horizon has been doubling approximately every seven months since 2019, though the trend may have accelerated in 2024. The increase in AI models' time horizons seems to be primarily driven by greater reliability and ability to adapt to mistakes, combined with better logical reasoning and tool use capabilities. We discuss the limitations of our results -- including their degree of external validity -- and the implications of increased autonomy for dangerous capabilities. If these results generalize to real-world software tasks, extrapolation of this trend predicts that within 5 years, AI systems will be capable of automating many software tasks that currently take humans a month. |
|
2025-03-19T00:00:00 | 2503.14492 | Cosmos-Transfer1: Conditional World Generation with Adaptive Multimodal Control | [
"NVIDIA",
"Hassan Abu Alhaija",
"Jose Alvarez",
"Maciej Bala",
"Tiffany Cai",
"Tianshi Cao",
"Liz Cha",
"Joshua Chen",
"Mike Chen",
"Francesco Ferroni",
"Sanja Fidler",
"Dieter Fox",
"Yunhao Ge",
"Jinwei Gu",
"Ali Hassani",
"Michael Isaev",
"Pooya Jannaty",
"Shiyi Lan",
"Tobias Lasser",
"Huan Ling",
"Ming-Yu Liu",
"Xian Liu",
"Yifan Lu",
"Alice Luo",
"Qianli Ma",
"Hanzi Mao",
"Fabio Ramos",
"Xuanchi Ren",
"Tianchang Shen",
"Shitao Tang",
"Ting-Chun Wang",
"Jay Wu",
"Jiashu Xu",
"Stella Xu",
"Kevin Xie",
"Yuchong Ye",
"Xiaodong Yang",
"Xiaohui Zeng",
"Yu Zeng"
] | https://github.com/nvidia-cosmos/cosmos-transfer1 | We introduce Cosmos-Transfer, a conditional world generation model that can generate world simulations based on multiple spatial control inputs of various modalities such as segmentation, depth, and edge. In the design, the spatial conditional scheme is adaptive and customizable. It allows weighting different conditional inputs differently at different spatial locations. This enables highly controllable world generation and finds use in various world-to-world transfer use cases, including Sim2Real. We conduct extensive evaluations to analyze the proposed model and demonstrate its applications for Physical AI, including robotics Sim2Real and autonomous vehicle data enrichment. We further demonstrate an inference scaling strategy to achieve real-time world generation with an NVIDIA GB200 NVL72 rack. To help accelerate research development in the field, we open-source our models and code at https://github.com/nvidia-cosmos/cosmos-transfer1. |
2025-03-19T00:00:00 | 2503.14478 | Creation-MMBench: Assessing Context-Aware Creative Intelligence in MLLM | [
"Xinyu Fang",
"Zhijian Chen",
"Kai Lan",
"Shengyuan Ding",
"Yingji Liang",
"Xiangyu Zhao",
"Farong Wen",
"Zicheng Zhang",
"Guofeng Zhang",
"Haodong Duan",
"Kai Chen",
"Dahua Lin"
] | https://github.com/open-compass/Creation-MMBench | Creativity is a fundamental aspect of intelligence, involving the ability to generate novel and appropriate solutions across diverse contexts. While Large Language Models (LLMs) have been extensively evaluated for their creative capabilities, the assessment of Multimodal Large Language Models (MLLMs) in this domain remains largely unexplored. To address this gap, we introduce Creation-MMBench, a multimodal benchmark specifically designed to evaluate the creative capabilities of MLLMs in real-world, image-based tasks. The benchmark comprises 765 test cases spanning 51 fine-grained tasks. To ensure rigorous evaluation, we define instance-specific evaluation criteria for each test case, guiding the assessment of both general response quality and factual consistency with visual inputs. Experimental results reveal that current open-source MLLMs significantly underperform compared to proprietary models in creative tasks. Furthermore, our analysis demonstrates that visual fine-tuning can negatively impact the base LLM's creative abilities. Creation-MMBench provides valuable insights for advancing MLLM creativity and establishes a foundation for future improvements in multimodal generative intelligence. Full data and evaluation code is released on https://github.com/open-compass/Creation-MMBench. |
2025-03-19T00:00:00 | 2503.12329 | CapArena: Benchmarking and Analyzing Detailed Image Captioning in the LLM Era | [
"Kanzhi Cheng",
"Wenpo Song",
"Jiaxin Fan",
"Zheng Ma",
"Qiushi Sun",
"Fangzhi Xu",
"Chenyang Yan",
"Nuo Chen",
"Jianbing Zhang",
"Jiajun Chen"
] | Image captioning has been a longstanding challenge in vision-language research. With the rise of LLMs, modern Vision-Language Models (VLMs) generate detailed and comprehensive image descriptions. However, benchmarking the quality of such captions remains unresolved. This paper addresses two key questions: (1) How well do current VLMs actually perform on image captioning, particularly compared to humans? We built CapArena, a platform with over 6000 pairwise caption battles and high-quality human preference votes. Our arena-style evaluation marks a milestone, showing that leading models like GPT-4o achieve or even surpass human performance, while most open-source models lag behind. (2) Can automated metrics reliably assess detailed caption quality? Using human annotations from CapArena, we evaluate traditional and recent captioning metrics, as well as VLM-as-a-Judge. Our analysis reveals that while some metrics (e.g., METEOR) show decent caption-level agreement with humans, their systematic biases lead to inconsistencies in model ranking. In contrast, VLM-as-a-Judge demonstrates robust discernment at both the caption and model levels. Building on these insights, we release CapArena-Auto, an accurate and efficient automated benchmark for detailed captioning, achieving 94.3% correlation with human rankings at just $4 per test. Data and resources will be open-sourced at https://caparena.github.io. |
|
2025-03-19T00:00:00 | 2503.12271 | Reflect-DiT: Inference-Time Scaling for Text-to-Image Diffusion Transformers via In-Context Reflection | [
"Shufan Li",
"Konstantinos Kallidromitis",
"Akash Gokul",
"Arsh Koneru",
"Yusuke Kato",
"Kazuki Kozuka",
"Aditya Grover"
] | The predominant approach to advancing text-to-image generation has been training-time scaling, where larger models are trained on more data using greater computational resources. While effective, this approach is computationally expensive, leading to growing interest in inference-time scaling to improve performance. Currently, inference-time scaling for text-to-image diffusion models is largely limited to best-of-N sampling, where multiple images are generated per prompt and a selection model chooses the best output. Inspired by the recent success of reasoning models like DeepSeek-R1 in the language domain, we introduce an alternative to naive best-of-N sampling by equipping text-to-image Diffusion Transformers with in-context reflection capabilities. We propose Reflect-DiT, a method that enables Diffusion Transformers to refine their generations using in-context examples of previously generated images alongside textual feedback describing necessary improvements. Instead of passively relying on random sampling and hoping for a better result in a future generation, Reflect-DiT explicitly tailors its generations to address specific aspects requiring enhancement. Experimental results demonstrate that Reflect-DiT improves performance on the GenEval benchmark (+0.19) using SANA-1.0-1.6B as a base model. Additionally, it achieves a new state-of-the-art score of 0.81 on GenEval while generating only 20 samples per prompt, surpassing the previous best score of 0.80, which was obtained using a significantly larger model (SANA-1.5-4.8B) with 2048 samples under the best-of-N approach. |
|
2025-03-19T00:00:00 | 2503.14495 | Temporal Consistency for LLM Reasoning Process Error Identification | [
"Jiacheng Guo",
"Yue Wu",
"Jiahao Qiu",
"Kaixuan Huang",
"Xinzhe Juan",
"Ling Yang",
"Mengdi Wang"
] | https://github.com/jcguo123/Temporal-Consistency | Verification is crucial for effective mathematical reasoning. We present a new temporal consistency method where verifiers iteratively refine their judgments based on the previous assessment. Unlike one-round verification or multi-model debate approaches, our method leverages consistency in a sequence of self-reflection actions to improve verification accuracy. Empirical evaluations across diverse mathematical process error identification benchmarks (Mathcheck, ProcessBench, and PRM800K) show consistent performance improvements over baseline methods. When applied to the recent DeepSeek R1 distilled models, our method demonstrates strong performance, enabling 7B/8B distilled models to outperform all 70B/72B models and GPT-4o on ProcessBench. Notably, the distilled 14B model with our method achieves performance comparable to Deepseek-R1. Our codes are available at https://github.com/jcguo123/Temporal-Consistency |
2025-03-19T00:00:00 | 2503.14504 | Aligning Multimodal LLM with Human Preference: A Survey | [
"Tao Yu",
"Yi-Fan Zhang",
"Chaoyou Fu",
"Junkang Wu",
"Jinda Lu",
"Kun Wang",
"Xingyu Lu",
"Yunhang Shen",
"Guibin Zhang",
"Dingjie Song",
"Yibo Yan",
"Tianlong Xu",
"Qingsong Wen",
"Zhang Zhang",
"Yan Huang",
"Liang Wang",
"Tieniu Tan"
] | https://github.com/BradyFU/Awesome-Multimodal-Large-Language-Models | Large language models (LLMs) can handle a wide variety of general tasks with simple prompts, without the need for task-specific training. Multimodal Large Language Models (MLLMs), built upon LLMs, have demonstrated impressive potential in tackling complex tasks involving visual, auditory, and textual data. However, critical issues related to truthfulness, safety, o1-like reasoning, and alignment with human preference remain insufficiently addressed. This gap has spurred the emergence of various alignment algorithms, each targeting different application scenarios and optimization goals. Recent studies have shown that alignment algorithms are a powerful approach to resolving the aforementioned challenges. In this paper, we aim to provide a comprehensive and systematic review of alignment algorithms for MLLMs. Specifically, we explore four key aspects: (1) the application scenarios covered by alignment algorithms, including general image understanding, multi-image, video, and audio, and extended multimodal applications; (2) the core factors in constructing alignment datasets, including data sources, model responses, and preference annotations; (3) the benchmarks used to evaluate alignment algorithms; and (4) a discussion of potential future directions for the development of alignment algorithms. This work seeks to help researchers organize current advancements in the field and inspire better alignment methods. The project page of this paper is available at https://github.com/BradyFU/Awesome-Multimodal-Large-Language-Models/tree/Alignment. |
2025-03-19T00:00:00 | 2503.14151 | Concat-ID: Towards Universal Identity-Preserving Video Synthesis | [
"Yong Zhong",
"Zhuoyi Yang",
"Jiayan Teng",
"Xiaotao Gu",
"Chongxuan Li"
] | We present Concat-ID, a unified framework for identity-preserving video generation. Concat-ID employs Variational Autoencoders to extract image features, which are concatenated with video latents along the sequence dimension, leveraging solely 3D self-attention mechanisms without the need for additional modules. A novel cross-video pairing strategy and a multi-stage training regimen are introduced to balance identity consistency and facial editability while enhancing video naturalness. Extensive experiments demonstrate Concat-ID's superiority over existing methods in both single and multi-identity generation, as well as its seamless scalability to multi-subject scenarios, including virtual try-on and background-controllable generation. Concat-ID establishes a new benchmark for identity-preserving video synthesis, providing a versatile and scalable solution for a wide range of applications. |
|
2025-03-19T00:00:00 | 2503.09443 | Florenz: Scaling Laws for Systematic Generalization in Vision-Language Models | [
"Julian Spravil",
"Sebastian Houben",
"Sven Behnke"
] | Cross-lingual transfer enables vision-language models (VLMs) to perform vision tasks in various languages with training data only in one language. Current approaches rely on large pre-trained multilingual language models. However, they face the curse of multilinguality, sacrificing downstream task performance for multilingual capabilities, struggling with lexical ambiguities, and falling behind recent advances. In this work, we study the scaling laws of systematic generalization with monolingual VLMs for multilingual tasks, focusing on the impact of model size and seen training samples. We propose Florenz, a monolingual encoder-decoder VLM with 0.4B to 11.2B parameters combining the pre-trained VLM Florence-2 and the large language model Gemma-2. Florenz is trained with varying compute budgets on a synthetic dataset that features intentionally incomplete language coverage for image captioning, thus, testing generalization from the fully covered translation task. We show that not only does indirectly learning unseen task-language pairs adhere to a scaling law, but also that with our data generation pipeline and the proposed Florenz model family, image captioning abilities can emerge in a specific language even when only data for the translation task is available. Fine-tuning on a mix of downstream datasets yields competitive performance and demonstrates promising scaling trends in multimodal machine translation (Multi30K, CoMMuTE), lexical disambiguation (CoMMuTE), and image captioning (Multi30K, XM3600, COCO Karpathy). |
|
2025-03-19T00:00:00 | 2503.13661 | Pensez: Less Data, Better Reasoning -- Rethinking French LLM | [
"Huy Hoang Ha"
] | Large language models (LLMs) have demonstrated remarkable capabilities in various natural language processing tasks. However, achieving strong performance in specialized domains like mathematical reasoning and non-English languages often requires extensive training on massive datasets. This paper investigates a contrasting approach: strategic fine-tuning on a small, high-quality, bilingual (English-French) dataset to enhance both the reasoning capabilities and French language proficiency of a large language model. Rather than relying on scale, we explore the hypothesis that targeted data curation and optimized training can achieve competitive, or even superior, performance. We demonstrate, through targeted supervised fine-tuning (SFT) on only 2,000 carefully selected samples, significant improvements in mathematical reasoning. Specifically, Pensez 7B exhibits an increase in accuracy of the base model up to 20% on the AIME25 and a 12% increase on a French MATH level 5 benchmark. These results challenge the prevailing assumption that massive datasets are aprerequisite for strong reasoning performance in LLMs, highlighting the potential of strategic data curation and optimized fine-tuning for enhancing both specialized skills and multilingual capabilities. Our findings have implications for the efficient development of high-performing, multilingual LLMs, especially in resource-constrained scenarios. |
|
2025-03-19T00:00:00 | 2503.12127 | Hyperbolic Safety-Aware Vision-Language Models | [
"Tobia Poppi",
"Tejaswi Kasarla",
"Pascal Mettes",
"Lorenzo Baraldi",
"Rita Cucchiara"
] | https://github.com/aimagelab/HySAC | Addressing the retrieval of unsafe content from vision-language models such as CLIP is an important step towards real-world integration. Current efforts have relied on unlearning techniques that try to erase the model's knowledge of unsafe concepts. While effective in reducing unwanted outputs, unlearning limits the model's capacity to discern between safe and unsafe content. In this work, we introduce a novel approach that shifts from unlearning to an awareness paradigm by leveraging the inherent hierarchical properties of the hyperbolic space. We propose to encode safe and unsafe content as an entailment hierarchy, where both are placed in different regions of hyperbolic space. Our HySAC, Hyperbolic Safety-Aware CLIP, employs entailment loss functions to model the hierarchical and asymmetrical relations between safe and unsafe image-text pairs. This modelling, ineffective in standard vision-language models due to their reliance on Euclidean embeddings, endows the model with awareness of unsafe content, enabling it to serve as both a multimodal unsafe classifier and a flexible content retriever, with the option to dynamically redirect unsafe queries toward safer alternatives or retain the original output. Extensive experiments show that our approach not only enhances safety recognition but also establishes a more adaptable and interpretable framework for content moderation in vision-language models. Our source code is available at https://github.com/aimagelab/HySAC. |
2025-03-19T00:00:00 | 2503.13265 | FlexWorld: Progressively Expanding 3D Scenes for Flexiable-View Synthesis | [
"Luxi Chen",
"Zihan Zhou",
"Min Zhao",
"Yikai Wang",
"Ge Zhang",
"Wenhao Huang",
"Hao Sun",
"Ji-Rong Wen",
"Chongxuan Li"
] | Generating flexible-view 3D scenes, including 360{\deg} rotation and zooming, from single images is challenging due to a lack of 3D data. To this end, we introduce FlexWorld, a novel framework consisting of two key components: (1) a strong video-to-video (V2V) diffusion model to generate high-quality novel view images from incomplete input rendered from a coarse scene, and (2) a progressive expansion process to construct a complete 3D scene. In particular, leveraging an advanced pre-trained video model and accurate depth-estimated training pairs, our V2V model can generate novel views under large camera pose variations. Building upon it, FlexWorld progressively generates new 3D content and integrates it into the global scene through geometry-aware scene fusion. Extensive experiments demonstrate the effectiveness of FlexWorld in generating high-quality novel view videos and flexible-view 3D scenes from single images, achieving superior visual quality under multiple popular metrics and datasets compared to existing state-of-the-art methods. Qualitatively, we highlight that FlexWorld can generate high-fidelity scenes with flexible views like 360{\deg} rotations and zooming. Project page: https://ml-gsai.github.io/FlexWorld. |
|
2025-03-19T00:00:00 | 2503.13111 | MM-Spatial: Exploring 3D Spatial Understanding in Multimodal LLMs | [
"Erik Daxberger",
"Nina Wenzel",
"David Griffiths",
"Haiming Gang",
"Justin Lazarow",
"Gefen Kohavi",
"Kai Kang",
"Marcin Eichner",
"Yinfei Yang",
"Afshin Dehghan",
"Peter Grasch"
] | Multimodal large language models (MLLMs) excel at 2D visual understanding but remain limited in their ability to reason about 3D space. In this work, we leverage large-scale high-quality 3D scene data with open-set annotations to introduce 1) a novel supervised fine-tuning dataset and 2) a new evaluation benchmark, focused on indoor scenes. Our Cubify Anything VQA (CA-VQA) data covers diverse spatial tasks including spatial relationship prediction, metric size and distance estimation, and 3D grounding. We show that CA-VQA enables us to train MM-Spatial, a strong generalist MLLM that also achieves state-of-the-art performance on 3D spatial understanding benchmarks, including our own. We show how incorporating metric depth and multi-view inputs (provided in CA-VQA) can further improve 3D understanding, and demonstrate that data alone allows our model to achieve depth perception capabilities comparable to dedicated monocular depth estimation models. We will publish our SFT dataset and benchmark. |
|
2025-03-19T00:00:00 | 2503.12303 | Towards Self-Improving Systematic Cognition for Next-Generation Foundation MLLMs | [
"Xiaoying Zhang",
"Da Peng",
"Yipeng Zhang",
"Zonghao Guo",
"Chengyue Wu",
"Chi Chen",
"Wei Ke",
"Helen Meng",
"Maosong Sun"
] | Despite their impressive capabilities, Multimodal Large Language Models (MLLMs) face challenges with fine-grained perception and complex reasoning. Prevalent multimodal pre-training approaches focus on enhancing perception by training on high-quality image captions due to the extremely high cost of collecting chain-of-thought (CoT) reasoning data for improving reasoning. While leveraging advanced MLLMs for caption generation enhances scalability, the outputs often lack comprehensiveness and accuracy. In this paper, we introduce Self-Improving cognition (SIcog), a self-learning framework designed to construct next-generation foundation MLLMs by enhancing their systematic cognitive capabilities through multimodal pre-training with self-generated data. Specifically, we propose Chain-of-Description, an approach that improves an MLLM's systematic perception by enabling step-by-step visual understanding, ensuring greater comprehensiveness and accuracy. Additionally, we adopt a structured CoT reasoning technique to enable MLLMs to integrate in-depth multimodal reasoning. To construct a next-generation foundation MLLM with self-improved cognition, SIcog first equips an MLLM with systematic perception and reasoning abilities using minimal external annotations. The enhanced models then generate detailed captions and CoT reasoning data, which are further curated through self-consistency. This curated data is ultimately used for multimodal pre-training to develop next-generation foundation models. Extensive experiments on both low- and high-resolution MLLMs across diverse benchmarks demonstrate that, with merely 213K self-generated pre-training samples, SIcog produces next-generation foundation MLLMs with significantly improved cognition, achieving benchmark-leading performance compared to prevalent pre-training approaches. |
|
2025-03-19T00:00:00 | 2503.14002 | MeshFleet: Filtered and Annotated 3D Vehicle Dataset for Domain Specific Generative Modeling | [
"Damian Boborzi",
"Phillip Mueller",
"Jonas Emrich",
"Dominik Schmid",
"Sebastian Mueller",
"Lars Mikelsons"
] | Generative models have recently made remarkable progress in the field of 3D objects. However, their practical application in fields like engineering remains limited since they fail to deliver the accuracy, quality, and controllability needed for domain-specific tasks. Fine-tuning large generative models is a promising perspective for making these models available in these fields. Creating high-quality, domain-specific 3D datasets is crucial for fine-tuning large generative models, yet the data filtering and annotation process remains a significant bottleneck. We present MeshFleet, a filtered and annotated 3D vehicle dataset extracted from Objaverse-XL, the most extensive publicly available collection of 3D objects. Our approach proposes a pipeline for automated data filtering based on a quality classifier. This classifier is trained on a manually labeled subset of Objaverse, incorporating DINOv2 and SigLIP embeddings, refined through caption-based analysis and uncertainty estimation. We demonstrate the efficacy of our filtering method through a comparative analysis against caption and image aesthetic score-based techniques and fine-tuning experiments with SV3D, highlighting the importance of targeted data selection for domain-specific 3D generative modeling. |
|
2025-03-19T00:00:00 | 2503.08893 | EvalTree: Profiling Language Model Weaknesses via Hierarchical Capability Trees | [
"Zhiyuan Zeng",
"Yizhong Wang",
"Hannaneh Hajishirzi",
"Pang Wei Koh"
] | An ideal model evaluation should achieve two goals: identifying where the model fails and providing actionable improvement guidance. Toward these goals for Language Model (LM) evaluations, we formulate the problem of generating a weakness profile, a set of weaknesses expressed in natural language, given an LM's performance on every individual instance in a benchmark. We introduce a suite of quantitative assessments to compare different weakness profiling methods. We also propose a weakness profiling method EvalTree. It constructs a capability tree where each node represents a capability described in natural language and is linked to a subset of benchmark instances that specifically evaluate this capability; it then extracts nodes where the LM performs poorly to generate a weakness profile. On the MATH and WildChat benchmarks, we show that EvalTree outperforms baseline weakness profiling methods by identifying weaknesses more precisely and comprehensively. Weakness profiling further enables weakness-guided data collection, and training data collection guided by EvalTree-identified weaknesses improves LM performance more than other data collection strategies. We also show how EvalTree exposes flaws in Chatbot Arena's human-voter-based evaluation practice. To facilitate future work, we release our code and an interface that allows practitioners to interactively explore the capability trees built by EvalTree. |
|
2025-03-19T00:00:00 | 2503.08683 | CoLMDriver: LLM-based Negotiation Benefits Cooperative Autonomous Driving | [
"Changxing Liu",
"Genjia Liu",
"Zijun Wang",
"Jinchang Yang",
"Siheng Chen"
] | https://github.com/cxliu0314/CoLMDriver | Vehicle-to-vehicle (V2V) cooperative autonomous driving holds great promise for improving safety by addressing the perception and prediction uncertainties inherent in single-agent systems. However, traditional cooperative methods are constrained by rigid collaboration protocols and limited generalization to unseen interactive scenarios. While LLM-based approaches offer generalized reasoning capabilities, their challenges in spatial planning and unstable inference latency hinder their direct application in cooperative driving. To address these limitations, we propose CoLMDriver, the first full-pipeline LLM-based cooperative driving system, enabling effective language-based negotiation and real-time driving control. CoLMDriver features a parallel driving pipeline with two key components: (i) an LLM-based negotiation module under an actor-critic paradigm, which continuously refines cooperation policies through feedback from previous decisions of all vehicles; and (ii) an intention-guided waypoint generator, which translates negotiation outcomes into executable waypoints. Additionally, we introduce InterDrive, a CARLA-based simulation benchmark comprising 10 challenging interactive driving scenarios for evaluating V2V cooperation. Experimental results demonstrate that CoLMDriver significantly outperforms existing approaches, achieving an 11% higher success rate across diverse highly interactive V2V driving scenarios. Code will be released on https://github.com/cxliu0314/CoLMDriver. |
2025-03-19T00:00:00 | 2503.10905 | Learning to Inference Adaptively for Multimodal Large Language Models | [
"Zhuoyan Xu",
"Khoi Duc Nguyen",
"Preeti Mukherjee",
"Saurabh Bagchi",
"Somali Chaterji",
"Yingyu Liang",
"Yin Li"
] | Multimodal Large Language Models (MLLMs) have shown impressive capabilities in reasoning, yet come with substantial computational cost, limiting their deployment in resource-constrained settings. Despite recent efforts on improving the efficiency of MLLMs, prior solutions fall short in responding to varying runtime conditions, in particular changing resource availability (e.g., contention due to the execution of other programs on the device). To bridge this gap, we introduce AdaLLaVA, an adaptive inference framework that learns to dynamically reconfigure operations in an MLLM during inference, accounting for the input data and a latency budget. We conduct extensive experiments across benchmarks involving question-answering, reasoning, and hallucination. Our results show that AdaLLaVA effectively adheres to input latency budget, achieving varying accuracy and latency tradeoffs at runtime. Further, we demonstrate that AdaLLaVA adapts to both input latency and content, can be integrated with token selection for enhanced efficiency, and generalizes across MLLMs. Our project webpage with code release is at https://zhuoyan-xu.github.io/ada-llava/. |
|
2025-03-19T00:00:00 | 2503.10522 | AudioX: Diffusion Transformer for Anything-to-Audio Generation | [
"Zeyue Tian",
"Yizhu Jin",
"Zhaoyang Liu",
"Ruibin Yuan",
"Xu Tan",
"Qifeng Chen",
"Wei Xue",
"Yike Guo"
] | Audio and music generation have emerged as crucial tasks in many applications, yet existing approaches face significant limitations: they operate in isolation without unified capabilities across modalities, suffer from scarce high-quality, multi-modal training data, and struggle to effectively integrate diverse inputs. In this work, we propose AudioX, a unified Diffusion Transformer model for Anything-to-Audio and Music Generation. Unlike previous domain-specific models, AudioX can generate both general audio and music with high quality, while offering flexible natural language control and seamless processing of various modalities including text, video, image, music, and audio. Its key innovation is a multi-modal masked training strategy that masks inputs across modalities and forces the model to learn from masked inputs, yielding robust and unified cross-modal representations. To address data scarcity, we curate two comprehensive datasets: vggsound-caps with 190K audio captions based on the VGGSound dataset, and V2M-caps with 6 million music captions derived from the V2M dataset. Extensive experiments demonstrate that AudioX not only matches or outperforms state-of-the-art specialized models, but also offers remarkable versatility in handling diverse input modalities and generation tasks within a unified architecture. The code and datasets will be available at https://zeyuet.github.io/AudioX/ |
|
2025-03-19T00:00:00 | 2503.12355 | Atlas: Multi-Scale Attention Improves Long Context Image Modeling | [
"Kumar Krishna Agrawal",
"Long Lian",
"Longchao Liu",
"Natalia Harguindeguy",
"Boyi Li",
"Alexander Bick",
"Maggie Chung",
"Trevor Darrell",
"Adam Yala"
] | https://github.com/yalalab/atlas | Efficiently modeling massive images is a long-standing challenge in machine learning. To this end, we introduce Multi-Scale Attention (MSA). MSA relies on two key ideas, (i) multi-scale representations (ii) bi-directional cross-scale communication. MSA creates O(log N) scales to represent the image across progressively coarser features and leverages cross-attention to propagate information across scales. We then introduce Atlas, a novel neural network architecture based on MSA. We demonstrate that Atlas significantly improves the compute-performance tradeoff of long-context image modeling in a high-resolution variant of ImageNet 100. At 1024px resolution, Atlas-B achieves 91.04% accuracy, comparable to ConvNext-B (91.92%) while being 4.3x faster. Atlas is 2.95x faster and 7.38% better than FasterViT, 2.25x faster and 4.96% better than LongViT. In comparisons against MambaVision-S, we find Atlas-S achieves 5%, 16% and 32% higher accuracy at 1024px, 2048px and 4096px respectively, while obtaining similar runtimes. Code for reproducing our experiments and pretrained models is available at https://github.com/yalalab/atlas. |
2025-03-19T00:00:00 | 2503.10284 | PyGDA: A Python Library for Graph Domain Adaptation | [
"Zhen Zhang",
"Meihan Liu",
"Bingsheng He"
] | https://github.com/pygda-team/pygda | Graph domain adaptation has emerged as a promising approach to facilitate knowledge transfer across different domains. Recently, numerous models have been proposed to enhance their generalization capabilities in this field. However, there is still no unified library that brings together existing techniques and simplifies their implementation. To fill this gap, we introduce PyGDA, an open-source Python library tailored for graph domain adaptation. As the first comprehensive library in this area, PyGDA covers more than 20 widely used graph domain adaptation methods together with different types of graph datasets. Specifically, PyGDA offers modular components, enabling users to seamlessly build custom models with a variety of commonly used utility functions. To handle large-scale graphs, PyGDA includes support for features such as sampling and mini-batch processing, ensuring efficient computation. In addition, PyGDA also includes comprehensive performance benchmarks and well-documented user-friendly API for both researchers and practitioners. To foster convenient accessibility, PyGDA is released under the MIT license at https://github.com/pygda-team/pygda, and the API documentation is https://pygda.readthedocs.io/en/stable/. |
2025-03-20T00:00:00 | 2503.15485 | TULIP: Towards Unified Language-Image Pretraining | [
"Zineng Tang",
"Long Lian",
"Seun Eisape",
"XuDong Wang",
"Roei Herzig",
"Adam Yala",
"Alane Suhr",
"Trevor Darrell",
"David M. Chan"
] | Despite the recent success of image-text contrastive models like CLIP and SigLIP, these models often struggle with vision-centric tasks that demand high-fidelity image understanding, such as counting, depth estimation, and fine-grained object recognition. These models, by performing language alignment, tend to prioritize high-level semantics over visual understanding, weakening their image understanding. On the other hand, vision-focused models are great at processing visual information but struggle to understand language, limiting their flexibility for language-driven tasks. In this work, we introduce TULIP, an open-source, drop-in replacement for existing CLIP-like models. Our method leverages generative data augmentation, enhanced image-image and text-text contrastive learning, and image/text reconstruction regularization to learn fine-grained visual features while preserving global semantic alignment. Our approach, scaling to over 1B parameters, outperforms existing state-of-the-art (SOTA) models across multiple benchmarks, establishing a new SOTA zero-shot performance on ImageNet-1K, delivering up to a 2times enhancement over SigLIP on RxRx1 in linear probing for few-shot classification, and improving vision-language models, achieving over 3times higher scores than SigLIP on MMVP. Our code/checkpoints are available at https://tulip-berkeley.github.io |
|
2025-03-20T00:00:00 | 2503.15475 | Cube: A Roblox View of 3D Intelligence | [
"Foundation AI Team",
"Kiran Bhat",
"Nishchaie Khanna",
"Karun Channa",
"Tinghui Zhou",
"Yiheng Zhu",
"Xiaoxia Sun",
"Charles Shang",
"Anirudh Sudarshan",
"Maurice Chu",
"Daiqing Li",
"Kangle Deng",
"Jean-Philippe Fauconnier",
"Tijmen Verhulsdonck",
"Maneesh Agrawala",
"Kayvon Fatahalian",
"Alexander Weiss",
"Christian Reiser",
"Ravi Kiran Chirravuri",
"Ravali Kandur",
"Alejandro Pelaez",
"Akash Garg",
"Michael Palleschi",
"Jessica Wang",
"Skylar Litz",
"Leon Liu",
"Anying Li",
"David Harmon",
"Derek Liu",
"Liangjun Feng",
"Denis Goupil",
"Lukas Kuczynski",
"Jihyun Yoon",
"Naveen Marri",
"Peiye Zhuang",
"Yinan Zhang",
"Brian Yin",
"Haomiao Jiang",
"Marcel van Workum",
"Thomas Lane",
"Bryce Erickson",
"Salil Pathare",
"Kyle Price",
"Anupam Singh",
"David Baszucki"
] | Foundation models trained on vast amounts of data have demonstrated remarkable reasoning and generation capabilities in the domains of text, images, audio and video. Our goal at Roblox is to build such a foundation model for 3D intelligence, a model that can support developers in producing all aspects of a Roblox experience, from generating 3D objects and scenes to rigging characters for animation to producing programmatic scripts describing object behaviors. We discuss three key design requirements for such a 3D foundation model and then present our first step towards building such a model. We expect that 3D geometric shapes will be a core data type and describe our solution for 3D shape tokenizer. We show how our tokenization scheme can be used in applications for text-to-shape generation, shape-to-text generation and text-to-scene generation. We demonstrate how these applications can collaborate with existing large language models (LLMs) to perform scene analysis and reasoning. We conclude with a discussion outlining our path to building a fully unified foundation model for 3D intelligence. |
|
2025-03-20T00:00:00 | 2503.12769 | ViSpeak: Visual Instruction Feedback in Streaming Videos | [
"Shenghao Fu",
"Qize Yang",
"Yuan-Ming Li",
"Yi-Xing Peng",
"Kun-Yu Lin",
"Xihan Wei",
"Jian-Fang Hu",
"Xiaohua Xie",
"Wei-Shi Zheng"
] | Recent advances in Large Multi-modal Models (LMMs) are primarily focused on offline video understanding. Instead, streaming video understanding poses great challenges to recent models due to its time-sensitive, omni-modal and interactive characteristics. In this work, we aim to extend the streaming video understanding from a new perspective and propose a novel task named Visual Instruction Feedback in which models should be aware of visual contents and learn to extract instructions from them. For example, when users wave their hands to agents, agents should recognize the gesture and start conversations with welcome information. Thus, following instructions in visual modality greatly enhances user-agent interactions. To facilitate research, we define seven key subtasks highly relevant to visual modality and collect the ViSpeak-Instruct dataset for training and the ViSpeak-Bench for evaluation. Further, we propose the ViSpeak model, which is a SOTA streaming video understanding LMM with GPT-4o-level performance on various streaming video understanding benchmarks. After finetuning on our ViSpeak-Instruct dataset, ViSpeak is equipped with basic visual instruction feedback ability, serving as a solid baseline for future research. |
|
2025-03-20T00:00:00 | 2503.15417 | Temporal Regularization Makes Your Video Generator Stronger | [
"Harold Haodong Chen",
"Haojian Huang",
"Xianfeng Wu",
"Yexin Liu",
"Yajing Bai",
"Wen-Jie Shu",
"Harry Yang",
"Ser-Nam Lim"
] | Temporal quality is a critical aspect of video generation, as it ensures consistent motion and realistic dynamics across frames. However, achieving high temporal coherence and diversity remains challenging. In this work, we explore temporal augmentation in video generation for the first time, and introduce FluxFlow for initial investigation, a strategy designed to enhance temporal quality. Operating at the data level, FluxFlow applies controlled temporal perturbations without requiring architectural modifications. Extensive experiments on UCF-101 and VBench benchmarks demonstrate that FluxFlow significantly improves temporal coherence and diversity across various video generation models, including U-Net, DiT, and AR-based architectures, while preserving spatial fidelity. These findings highlight the potential of temporal augmentation as a simple yet effective approach to advancing video generation quality. |
|
2025-03-20T00:00:00 | 2503.12532 | STEVE: AStep Verification Pipeline for Computer-use Agent Training | [
"Fanbin Lu",
"Zhisheng Zhong",
"Ziqin Wei",
"Shu Liu",
"Chi-Wing Fu",
"Jiaya Jia"
] | https://github.com/FanbinLu/STEVE | Developing AI agents to autonomously manipulate graphical user interfaces is a long challenging task. Recent advances in data scaling law inspire us to train computer-use agents with a scaled instruction set, yet using behavior cloning to train agents still requires immense high-quality trajectories. To meet the scalability need, we designed STEVE, a step verification pipeline for computer-use agent training. First, we establish a large instruction set for computer-use agents and collect trajectory data with some suboptimal agents. GPT-4o is used to verify the correctness of each step in the trajectories based on the screens before and after the action execution, assigning each step with a binary label. Last, we adopt the Kahneman and Tversky Optimization to optimize the agent from the binary stepwise labels. Extensive experiments manifest that our agent outperforms supervised finetuning by leveraging both positive and negative actions within a trajectory. Also, STEVE enables us to train a 7B vision-language model as a computer-use agent, achieving leading performance in the challenging live desktop environment WinAgentArena with great efficiency at a reduced cost. Code and data: https://github.com/FanbinLu/STEVE. |
2025-03-20T00:00:00 | 2503.14505 | MusicInfuser: Making Video Diffusion Listen and Dance | [
"Susung Hong",
"Ira Kemelmacher-Shlizerman",
"Brian Curless",
"Steven M. Seitz"
] | We introduce MusicInfuser, an approach for generating high-quality dance videos that are synchronized to a specified music track. Rather than attempting to design and train a new multimodal audio-video model, we show how existing video diffusion models can be adapted to align with musical inputs by introducing lightweight music-video cross-attention and a low-rank adapter. Unlike prior work requiring motion capture data, our approach fine-tunes only on dance videos. MusicInfuser achieves high-quality music-driven video generation while preserving the flexibility and generative capabilities of the underlying models. We introduce an evaluation framework using Video-LLMs to assess multiple dimensions of dance generation quality. The project page and code are available at https://susunghong.github.io/MusicInfuser. |
|
2025-03-20T00:00:00 | 2503.15055 | ELTEX: A Framework for Domain-Driven Synthetic Data Generation | [
"Arina Razmyslovich",
"Kseniia Murasheva",
"Sofia Sedlova",
"Julien Capitaine",
"Eugene Dmitriev"
] | We present ELTEX (Efficient LLM Token Extraction), a domain-driven framework for generating high-quality synthetic training data in specialized domains. While Large Language Models (LLMs) have shown impressive general capabilities, their performance in specialized domains like cybersecurity remains limited by the scarcity of domain-specific training data. ELTEX addresses this challenge by systematically integrating explicit domain indicator extraction with dynamic prompting to preserve critical domain knowledge throughout the generation process. We demonstrate ELTEX's effectiveness in the context of blockchain-related cyberattack detection, where we fine-tune Gemma-2B using various combinations of real and ELTEX-generated data. Our results show that the ELTEX-enhanced model achieves performance competitive with GPT-4 across both standard classification metrics and uncertainty calibration, while requiring significantly fewer computational resources. We release a curated synthetic dataset of social media texts for cyberattack detection in blockchain. Our work demonstrates that domain-driven synthetic data generation can effectively bridge the performance gap between resource-efficient models and larger architectures in specialized domains. |
|
2025-03-20T00:00:00 | 2503.14868 | Efficient Personalization of Quantized Diffusion Model without Backpropagation | [
"Hoigi Seo",
"Wongi Jeong",
"Kyungryeol Lee",
"Se Young Chun"
] | Diffusion models have shown remarkable performance in image synthesis, but they demand extensive computational and memory resources for training, fine-tuning and inference. Although advanced quantization techniques have successfully minimized memory usage for inference, training and fine-tuning these quantized models still require large memory possibly due to dequantization for accurate computation of gradients and/or backpropagation for gradient-based algorithms. However, memory-efficient fine-tuning is particularly desirable for applications such as personalization that often must be run on edge devices like mobile phones with private data. In this work, we address this challenge by quantizing a diffusion model with personalization via Textual Inversion and by leveraging a zeroth-order optimization on personalization tokens without dequantization so that it does not require gradient and activation storage for backpropagation that consumes considerable memory. Since a gradient estimation using zeroth-order optimization is quite noisy for a single or a few images in personalization, we propose to denoise the estimated gradient by projecting it onto a subspace that is constructed with the past history of the tokens, dubbed Subspace Gradient. In addition, we investigated the influence of text embedding in image generation, leading to our proposed time steps sampling, dubbed Partial Uniform Timestep Sampling for sampling with effective diffusion timesteps. Our method achieves comparable performance to prior methods in image and text alignment scores for personalizing Stable Diffusion with only forward passes while reducing training memory demand up to 8.2times. |
|
2025-03-20T00:00:00 | 2503.15265 | DeepMesh: Auto-Regressive Artist-mesh Creation with Reinforcement Learning | [
"Ruowen Zhao",
"Junliang Ye",
"Zhengyi Wang",
"Guangce Liu",
"Yiwen Chen",
"Yikai Wang",
"Jun Zhu"
] | Triangle meshes play a crucial role in 3D applications for efficient manipulation and rendering. While auto-regressive methods generate structured meshes by predicting discrete vertex tokens, they are often constrained by limited face counts and mesh incompleteness. To address these challenges, we propose DeepMesh, a framework that optimizes mesh generation through two key innovations: (1) an efficient pre-training strategy incorporating a novel tokenization algorithm, along with improvements in data curation and processing, and (2) the introduction of Reinforcement Learning (RL) into 3D mesh generation to achieve human preference alignment via Direct Preference Optimization (DPO). We design a scoring standard that combines human evaluation with 3D metrics to collect preference pairs for DPO, ensuring both visual appeal and geometric accuracy. Conditioned on point clouds and images, DeepMesh generates meshes with intricate details and precise topology, outperforming state-of-the-art methods in both precision and quality. Project page: https://zhaorw02.github.io/DeepMesh/ |
|
2025-03-20T00:00:00 | 2503.13360 | Mitigating Visual Forgetting via Take-along Visual Conditioning for Multi-modal Long CoT Reasoning | [
"Hai-Long Sun",
"Zhun Sun",
"Houwen Peng",
"Han-Jia Ye"
] | Recent advancements in Large Language Models (LLMs) have demonstrated enhanced reasoning capabilities, evolving from Chain-of-Thought (CoT) prompting to advanced, product-oriented solutions like OpenAI o1. During our re-implementation of this model, we noticed that in multimodal tasks requiring visual input (e.g., geometry problems), Multimodal LLMs (MLLMs) struggle to maintain focus on the visual information, in other words, MLLMs suffer from a gradual decline in attention to visual information as reasoning progresses, causing text-over-relied outputs. To investigate this, we ablate image inputs during long-chain reasoning. Concretely, we truncate the reasoning process midway, then re-complete the reasoning process with the input image removed. We observe only a ~2% accuracy drop on MathVista's test-hard subset, revealing the model's textual outputs dominate the following reasoning process. Motivated by this, we propose Take-along Visual Conditioning (TVC), a strategy that shifts image input to critical reasoning stages and compresses redundant visual tokens via dynamic pruning. This methodology helps the model retain attention to the visual components throughout the reasoning. Our approach achieves state-of-the-art performance on average across five mathematical reasoning benchmarks (+3.4% vs previous sota), demonstrating the effectiveness of TVC in enhancing multimodal reasoning systems. |
|
2025-03-20T00:00:00 | 2503.13288 | φ-Decoding: Adaptive Foresight Sampling for Balanced Inference-Time Exploration and Exploitation | [
"Fangzhi Xu",
"Hang Yan",
"Chang Ma",
"Haiteng Zhao",
"Jun Liu",
"Qika Lin",
"Zhiyong Wu"
] | https://github.com/xufangzhi/phi-Decoding | Inference-time optimization scales computation to derive deliberate reasoning steps for effective performance. While previous search-based strategies address the short-sightedness of auto-regressive generation, the vast search space leads to excessive exploration and insufficient exploitation. To strike an efficient balance to derive the optimal step, we frame the decoding strategy as foresight sampling, leveraging simulated future steps to obtain globally optimal step estimation. Built on it, we propose a novel decoding strategy, named phi-Decoding. To provide a precise and expressive estimation of step value, phi-Decoding approximates two distributions via foresight and clustering. Sampling from the joint distribution, the optimal steps can be selected for exploitation. To support adaptive computation allocation, we propose in-width and in-depth pruning strategies, featuring a light-weight solution to achieve inference efficiency. Extensive experiments across seven benchmarks show phi-Decoding outperforms strong baselines in both performance and efficiency. Additional analysis demonstrates its generalization across various LLMs and scalability across a wide range of computing budgets. The code will be released at https://github.com/xufangzhi/phi-Decoding, and the open-source PyPI package is coming soon. |
2025-03-20T00:00:00 | 2503.15264 | LEGION: Learning to Ground and Explain for Synthetic Image Detection | [
"Hengrui Kang",
"Siwei Wen",
"Zichen Wen",
"Junyan Ye",
"Weijia Li",
"Peilin Feng",
"Baichuan Zhou",
"Bin Wang",
"Dahua Lin",
"Linfeng Zhang",
"Conghui He"
] | The rapid advancements in generative technology have emerged as a double-edged sword. While offering powerful tools that enhance convenience, they also pose significant social concerns. As defenders, current synthetic image detection methods often lack artifact-level textual interpretability and are overly focused on image manipulation detection, and current datasets usually suffer from outdated generators and a lack of fine-grained annotations. In this paper, we introduce SynthScars, a high-quality and diverse dataset consisting of 12,236 fully synthetic images with human-expert annotations. It features 4 distinct image content types, 3 categories of artifacts, and fine-grained annotations covering pixel-level segmentation, detailed textual explanations, and artifact category labels. Furthermore, we propose LEGION (LEarning to Ground and explain for Synthetic Image detectiON), a multimodal large language model (MLLM)-based image forgery analysis framework that integrates artifact detection, segmentation, and explanation. Building upon this capability, we further explore LEGION as a controller, integrating it into image refinement pipelines to guide the generation of higher-quality and more realistic images. Extensive experiments show that LEGION outperforms existing methods across multiple benchmarks, particularly surpassing the second-best traditional expert on SynthScars by 3.31% in mIoU and 7.75% in F1 score. Moreover, the refined images generated under its guidance exhibit stronger alignment with human preferences. The code, model, and dataset will be released. |
|
2025-03-20T00:00:00 | 2503.11227 | GKG-LLM: A Unified Framework for Generalized Knowledge Graph Construction | [
"Jian Zhang",
"Bifan Wei",
"Shihao Qi",
"haiping Zhu",
"Jun Liu",
"Qika Lin"
] | The construction of Generalized Knowledge Graph (GKG), including knowledge graph, event knowledge graph and commonsense knowledge graph, is fundamental for various natural language processing tasks. Current studies typically construct these types of graph separately, overlooking holistic insights and potential unification that could be beneficial in computing resources and usage perspectives. However, a key challenge in developing a unified framework for GKG is obstacles arising from task-specific differences. In this study, we propose a unified framework for constructing generalized knowledge graphs to address this challenge. First, we collect data from 15 sub-tasks in 29 datasets across the three types of graphs, categorizing them into in-sample, counter-task, and out-of-distribution (OOD) data. Then, we propose a three-stage curriculum learning fine-tuning framework, by iteratively injecting knowledge from the three types of graphs into the Large Language Models. Extensive experiments show that our proposed model improves the construction of all three graph types across in-domain, OOD and counter-task data. |
|
2025-03-20T00:00:00 | 2503.12963 | Unlock Pose Diversity: Accurate and Efficient Implicit Keypoint-based Spatiotemporal Diffusion for Audio-driven Talking Portrait | [
"Chaolong Yang",
"Kai Yao",
"Yuyao Yan",
"Chenru Jiang",
"Weiguang Zhao",
"Jie Sun",
"Guangliang Cheng",
"Yifei Zhang",
"Bin Dong",
"Kaizhu Huang"
] | https://github.com/chaolongy/KDTalker | Audio-driven single-image talking portrait generation plays a crucial role in virtual reality, digital human creation, and filmmaking. Existing approaches are generally categorized into keypoint-based and image-based methods. Keypoint-based methods effectively preserve character identity but struggle to capture fine facial details due to the fixed points limitation of the 3D Morphable Model. Moreover, traditional generative networks face challenges in establishing causality between audio and keypoints on limited datasets, resulting in low pose diversity. In contrast, image-based approaches produce high-quality portraits with diverse details using the diffusion network but incur identity distortion and expensive computational costs. In this work, we propose KDTalker, the first framework to combine unsupervised implicit 3D keypoint with a spatiotemporal diffusion model. Leveraging unsupervised implicit 3D keypoints, KDTalker adapts facial information densities, allowing the diffusion process to model diverse head poses and capture fine facial details flexibly. The custom-designed spatiotemporal attention mechanism ensures accurate lip synchronization, producing temporally consistent, high-quality animations while enhancing computational efficiency. Experimental results demonstrate that KDTalker achieves state-of-the-art performance regarding lip synchronization accuracy, head pose diversity, and execution efficiency.Our codes are available at https://github.com/chaolongy/KDTalker. |
2025-03-20T00:00:00 | 2503.14891 | MetaLadder: Ascending Mathematical Solution Quality via Analogical-Problem Reasoning Transfer | [
"Honglin Lin",
"Zhuoshi Pan",
"Yu Li",
"Qizhi Pei",
"Xin Gao",
"Mengzhang Cai",
"Conghui He",
"Lijun Wu"
] | https://github.com/LHL3341/MetaLadder | Large Language Models (LLMs) have demonstrated promising capabilities in solving mathematical reasoning tasks, leveraging Chain-of-Thought (CoT) data as a vital component in guiding answer generation. Current paradigms typically generate CoT and answers directly for a given problem, diverging from human problem-solving strategies to some extent. Humans often solve problems by recalling analogous cases and leveraging their solutions to reason about the current task. Inspired by this cognitive process, we propose MetaLadder, a novel framework that explicitly prompts LLMs to recall and reflect on meta-problems, those structurally or semantically analogous problems, alongside their CoT solutions before addressing the target problem. Additionally, we introduce a problem-restating mechanism to enhance the model's comprehension of the target problem by regenerating the original question, which further improves reasoning accuracy. Therefore, the model can achieve reasoning transfer from analogical problems, mimicking human-like "learning from examples" and generalization abilities. Extensive experiments on mathematical benchmarks demonstrate that our MetaLadder significantly boosts LLMs' problem-solving accuracy, largely outperforming standard CoT-based methods (10.3\% accuracy gain) and other methods. Our code and data has been released at https://github.com/LHL3341/MetaLadder. |
2025-03-20T00:00:00 | 2503.13517 | CURIE: Evaluating LLMs On Multitask Scientific Long Context Understanding and Reasoning | [
"Hao Cui",
"Zahra Shamsi",
"Gowoon Cheon",
"Xuejian Ma",
"Shutong Li",
"Maria Tikhanovskaya",
"Peter Norgaard",
"Nayantara Mudur",
"Martyna Plomecka",
"Paul Raccuglia",
"Yasaman Bahri",
"Victor V. Albert",
"Pranesh Srinivasan",
"Haining Pan",
"Philippe Faist",
"Brian Rohr",
"Michael J. Statt",
"Dan Morris",
"Drew Purves",
"Elise Kleeman",
"Ruth Alcantara",
"Matthew Abraham",
"Muqthar Mohammad",
"Ean Phing VanLee",
"Chenfei Jiang",
"Elizabeth Dorfman",
"Eun-Ah Kim",
"Michael P Brenner",
"Viren Jain",
"Sameera Ponda",
"Subhashini Venugopalan"
] | https://github.com/google/curie | Scientific problem-solving involves synthesizing information while applying expert knowledge. We introduce CURIE, a scientific long-Context Understanding,Reasoning and Information Extraction benchmark to measure the potential of Large Language Models (LLMs) in scientific problem-solving and assisting scientists in realistic workflows. This benchmark introduces ten challenging tasks with a total of 580 problems and solution pairs curated by experts in six disciplines - materials science, condensed matter physics, quantum computing, geospatial analysis, biodiversity, and proteins - covering both experimental and theoretical work-flows in science. We evaluate a range of closed and open LLMs on tasks in CURIE which requires domain expertise, comprehension of long in-context information,and multi-step reasoning. While Gemini Flash 2.0 and Claude-3 show consistent high comprehension across domains, the popular GPT-4o and command-R+ fail dramatically on protein sequencing tasks. With the best performance at 32% there is much room for improvement for all models. We hope that insights gained from CURIE can guide the future development of LLMs in sciences. Evaluation code and data are in https://github.com/google/curie |
2025-03-20T00:00:00 | 2503.11557 | VERIFY: A Benchmark of Visual Explanation and Reasoning for Investigating Multimodal Reasoning Fidelity | [
"Jing Bi",
"Junjia Guo",
"Susan Liang",
"Guangyu Sun",
"Luchuan Song",
"Yunlong Tang",
"Jinxi He",
"Jiarui Wu",
"Ali Vosoughi",
"Chen Chen",
"Chenliang Xu"
] | Visual reasoning is central to human cognition, enabling individuals to interpret and abstractly understand their environment. Although recent Multimodal Large Language Models (MLLMs) have demonstrated impressive performance across language and vision-language tasks, existing benchmarks primarily measure recognition-based skills and inadequately assess true visual reasoning capabilities. To bridge this critical gap, we introduce VERIFY, a benchmark explicitly designed to isolate and rigorously evaluate the visual reasoning capabilities of state-of-the-art MLLMs. VERIFY compels models to reason primarily from visual information, providing minimal textual context to reduce reliance on domain-specific knowledge and linguistic biases. Each problem is accompanied by a human-annotated reasoning path, making it the first to provide in-depth evaluation of model decision-making processes. Additionally, we propose novel metrics that assess visual reasoning fidelity beyond mere accuracy, highlighting critical imbalances in current model reasoning patterns. Our comprehensive benchmarking of leading MLLMs uncovers significant limitations, underscoring the need for a balanced and holistic approach to both perception and reasoning. For more teaser and testing, visit our project page (https://verify-eqh.pages.dev/). |
|
2025-03-20T00:00:00 | 2503.15354 | Optimizing Decomposition for Optimal Claim Verification | [
"Yining Lu",
"Noah Ziems",
"Hy Dang",
"Meng Jiang"
] | Current research on the Decompose-Then-Verify paradigm for evaluating the factuality of long-form text typically treats decomposition and verification in isolation, overlooking their interactions and potential misalignment. We find that existing decomposition policies, typically hand-crafted demonstrations, do not align well with downstream verifiers in terms of atomicity -- a novel metric quantifying information density -- leading to suboptimal verification results. We formulate finding the optimal decomposition policy for optimal verification as a bilevel optimization problem. To approximate a solution for this strongly NP-hard problem, we propose dynamic decomposition, a reinforcement learning framework that leverages verifier feedback to learn a policy for dynamically decomposing claims to verifier-preferred atomicity. Experimental results show that dynamic decomposition outperforms existing decomposition policies, improving verification confidence by 0.07 and accuracy by 0.12 (on a 0-1 scale) on average across varying verifiers, datasets, and atomcities of input claims. |
|
2025-03-20T00:00:00 | 2503.14830 | Decompositional Neural Scene Reconstruction with Generative Diffusion Prior | [
"Junfeng Ni",
"Yu Liu",
"Ruijie Lu",
"Zirui Zhou",
"Song-Chun Zhu",
"Yixin Chen",
"Siyuan Huang"
] | Decompositional reconstruction of 3D scenes, with complete shapes and detailed texture of all objects within, is intriguing for downstream applications but remains challenging, particularly with sparse views as input. Recent approaches incorporate semantic or geometric regularization to address this issue, but they suffer significant degradation in underconstrained areas and fail to recover occluded regions. We argue that the key to solving this problem lies in supplementing missing information for these areas. To this end, we propose DP-Recon, which employs diffusion priors in the form of Score Distillation Sampling (SDS) to optimize the neural representation of each individual object under novel views. This provides additional information for the underconstrained areas, but directly incorporating diffusion prior raises potential conflicts between the reconstruction and generative guidance. Therefore, we further introduce a visibility-guided approach to dynamically adjust the per-pixel SDS loss weights. Together these components enhance both geometry and appearance recovery while remaining faithful to input images. Extensive experiments across Replica and ScanNet++ demonstrate that our method significantly outperforms SOTA methods. Notably, it achieves better object reconstruction under 10 views than the baselines under 100 views. Our method enables seamless text-based editing for geometry and appearance through SDS optimization and produces decomposed object meshes with detailed UV maps that support photorealistic Visual effects (VFX) editing. The project page is available at https://dp-recon.github.io/. |
|
2025-03-20T00:00:00 | 2503.15450 | SkyLadder: Better and Faster Pretraining via Context Window Scheduling | [
"Tongyao Zhu",
"Qian Liu",
"Haonan Wang",
"Shiqi Chen",
"Xiangming Gu",
"Tianyu Pang",
"Min-Yen Kan"
] | https://github.com/sail-sg/SkyLadder | Recent advancements in LLM pretraining have featured ever-expanding context windows to process longer sequences. However, our pilot study reveals that models pretrained with shorter context windows consistently outperform their long-context counterparts under a fixed token budget. This finding motivates us to explore an optimal context window scheduling strategy to better balance long-context capability with pretraining efficiency. To this end, we propose SkyLadder, a simple yet effective approach that implements a short-to-long context window transition. SkyLadder preserves strong standard benchmark performance, while matching or exceeding baseline results on long context tasks. Through extensive experiments, we pre-train 1B-parameter models (up to 32K context) and 3B-parameter models (8K context) on 100B tokens, demonstrating that SkyLadder yields consistent gains of up to 3.7% on common benchmarks, while achieving up to 22% faster training speeds compared to baselines. The code is at https://github.com/sail-sg/SkyLadder. |
2025-03-20T00:00:00 | 2503.14434 | LLM-FE: Automated Feature Engineering for Tabular Data with LLMs as Evolutionary Optimizers | [
"Nikhil Abhyankar",
"Parshin Shojaee",
"Chandan K. Reddy"
] | Automated feature engineering plays a critical role in improving predictive model performance for tabular learning tasks. Traditional automated feature engineering methods are limited by their reliance on pre-defined transformations within fixed, manually designed search spaces, often neglecting domain knowledge. Recent advances using Large Language Models (LLMs) have enabled the integration of domain knowledge into the feature engineering process. However, existing LLM-based approaches use direct prompting or rely solely on validation scores for feature selection, failing to leverage insights from prior feature discovery experiments or establish meaningful reasoning between feature generation and data-driven performance. To address these challenges, we propose LLM-FE, a novel framework that combines evolutionary search with the domain knowledge and reasoning capabilities of LLMs to automatically discover effective features for tabular learning tasks. LLM-FE formulates feature engineering as a program search problem, where LLMs propose new feature transformation programs iteratively, and data-driven feedback guides the search process. Our results demonstrate that LLM-FE consistently outperforms state-of-the-art baselines, significantly enhancing the performance of tabular prediction models across diverse classification and regression benchmarks. |
|
2025-03-20T00:00:00 | 2503.15478 | SWEET-RL: Training Multi-Turn LLM Agents on Collaborative Reasoning Tasks | [
"Yifei Zhou",
"Song Jiang",
"Yuandong Tian",
"Jason Weston",
"Sergey Levine",
"Sainbayar Sukhbaatar",
"Xian Li"
] | Large language model (LLM) agents need to perform multi-turn interactions in real-world tasks. However, existing multi-turn RL algorithms for optimizing LLM agents fail to perform effective credit assignment over multiple turns while leveraging the generalization capabilities of LLMs and it remains unclear how to develop such algorithms. To study this, we first introduce a new benchmark, ColBench, where an LLM agent interacts with a human collaborator over multiple turns to solve realistic tasks in backend programming and frontend design. Building on this benchmark, we propose a novel RL algorithm, SWEET-RL (RL with Step-WisE Evaluation from Training-time information), that uses a carefully designed optimization objective to train a critic model with access to additional training-time information. The critic provides step-level rewards for improving the policy model. Our experiments demonstrate that SWEET-RL achieves a 6% absolute improvement in success and win rates on ColBench compared to other state-of-the-art multi-turn RL algorithms, enabling Llama-3.1-8B to match or exceed the performance of GPT4-o in realistic collaborative content creation. |
|
2025-03-20T00:00:00 | 2503.13553 | LLM-Mediated Guidance of MARL Systems | [
"Philipp D. Siedler",
"Ian Gemp"
] | In complex multi-agent environments, achieving efficient learning and desirable behaviours is a significant challenge for Multi-Agent Reinforcement Learning (MARL) systems. This work explores the potential of combining MARL with Large Language Model (LLM)-mediated interventions to guide agents toward more desirable behaviours. Specifically, we investigate how LLMs can be used to interpret and facilitate interventions that shape the learning trajectories of multiple agents. We experimented with two types of interventions, referred to as controllers: a Natural Language (NL) Controller and a Rule-Based (RB) Controller. The NL Controller, which uses an LLM to simulate human-like interventions, showed a stronger impact than the RB Controller. Our findings indicate that agents particularly benefit from early interventions, leading to more efficient training and higher performance. Both intervention types outperform the baseline without interventions, highlighting the potential of LLM-mediated guidance to accelerate training and enhance MARL performance in challenging environments. |
|
2025-03-21T00:00:00 | 2503.16356 | CaKE: Circuit-aware Editing Enables Generalizable Knowledge Learners | [
"Yunzhi Yao",
"Jizhan Fang",
"Jia-Chen Gu",
"Ningyu Zhang",
"Shumin Deng",
"Huajun Chen",
"Nanyun Peng"
] | https://github.com/zjunlp/CaKE | Knowledge Editing (KE) enables the modification of outdated or incorrect information in large language models (LLMs). While existing KE methods can update isolated facts, they struggle to generalize these updates to multi-hop reasoning tasks that depend on the modified knowledge. Through an analysis of reasoning circuits -- the neural pathways LLMs use for knowledge-based inference, we observe that current layer-localized KE approaches, such as MEMIT and WISE, which edit only single or a few model layers, struggle to effectively incorporate updated information into these reasoning pathways. To address this limitation, we propose CaKE (Circuit-aware Knowledge Editing), a novel method that enables more effective integration of updated knowledge in LLMs. CaKE leverages strategically curated data, guided by our circuits-based analysis, that enforces the model to utilize the modified knowledge, stimulating the model to develop appropriate reasoning circuits for newly integrated knowledge. Experimental results show that CaKE enables more accurate and consistent use of updated knowledge across related reasoning tasks, leading to an average of 20% improvement in multi-hop reasoning accuracy on MQuAKE dataset compared to existing KE methods. We release the code and data in https://github.com/zjunlp/CaKE. |
2025-03-21T00:00:00 | 2503.16278 | Uni-3DAR: Unified 3D Generation and Understanding via Autoregression on Compressed Spatial Tokens | [
"Shuqi Lu",
"Haowei Lin",
"Lin Yao",
"Zhifeng Gao",
"Xiaohong Ji",
"Weinan E",
"Linfeng Zhang",
"Guolin Ke"
] | https://github.com/dptech-corp/Uni-3DAR | Recent advancements in large language models and their multi-modal extensions have demonstrated the effectiveness of unifying generation and understanding through autoregressive next-token prediction. However, despite the critical role of 3D structural generation and understanding ({3D GU}) in AI for science, these tasks have largely evolved independently, with autoregressive methods remaining underexplored. To bridge this gap, we introduce Uni-3DAR, a unified framework that seamlessly integrates {3D GU} tasks via autoregressive prediction. At its core, Uni-3DAR employs a novel hierarchical tokenization that compresses 3D space using an octree, leveraging the inherent sparsity of 3D structures. It then applies an additional tokenization for fine-grained structural details, capturing key attributes such as atom types and precise spatial coordinates in microscopic 3D structures. We further propose two optimizations to enhance efficiency and effectiveness. The first is a two-level subtree compression strategy, which reduces the octree token sequence by up to 8x. The second is a masked next-token prediction mechanism tailored for dynamically varying token positions, significantly boosting model performance. By combining these strategies, Uni-3DAR successfully unifies diverse {3D GU} tasks within a single autoregressive framework. Extensive experiments across multiple microscopic {3D GU} tasks, including molecules, proteins, polymers, and crystals, validate its effectiveness and versatility. Notably, Uni-3DAR surpasses previous state-of-the-art diffusion models by a substantial margin, achieving up to 256\% relative improvement while delivering inference speeds up to 21.8x faster. The code is publicly available at https://github.com/dptech-corp/Uni-3DAR. |
2025-03-21T00:00:00 | 2503.15567 | Towards Unified Latent Space for 3D Molecular Latent Diffusion Modeling | [
"Yanchen Luo",
"Zhiyuan Liu",
"Yi Zhao",
"Sihang Li",
"Kenji Kawaguchi",
"Tat-Seng Chua",
"Xiang Wang"
] | 3D molecule generation is crucial for drug discovery and material science, requiring models to process complex multi-modalities, including atom types, chemical bonds, and 3D coordinates. A key challenge is integrating these modalities of different shapes while maintaining SE(3) equivariance for 3D coordinates. To achieve this, existing approaches typically maintain separate latent spaces for invariant and equivariant modalities, reducing efficiency in both training and sampling. In this work, we propose Unified Variational Auto-Encoder for 3D Molecular Latent Diffusion Modeling (UAE-3D), a multi-modal VAE that compresses 3D molecules into latent sequences from a unified latent space, while maintaining near-zero reconstruction error. This unified latent space eliminates the complexities of handling multi-modality and equivariance when performing latent diffusion modeling. We demonstrate this by employing the Diffusion Transformer--a general-purpose diffusion model without any molecular inductive bias--for latent generation. Extensive experiments on GEOM-Drugs and QM9 datasets demonstrate that our method significantly establishes new benchmarks in both de novo and conditional 3D molecule generation, achieving leading efficiency and quality. |
|
2025-03-21T00:00:00 | 2503.16252 | Fin-R1: A Large Language Model for Financial Reasoning through Reinforcement Learning | [
"Zhaowei Liu",
"Xin Guo",
"Fangqi Lou",
"Lingfeng Zeng",
"Jinyi Niu",
"Zixuan Wang",
"Jiajie Xu",
"Weige Cai",
"Ziwei Yang",
"Xueqian Zhao",
"Chao Li",
"Sheng Xu",
"Dezhi Chen",
"Yun Chen",
"Zuo Bai",
"Liwen Zhang"
] | https://github.com/SUFE-AIFLM-Lab/Fin-R1 | Reasoning large language models are rapidly evolving across various domains. However, their capabilities in handling complex financial tasks still require in-depth exploration. In this paper, we introduce Fin-R1, a reasoning large language model specifically designed for the financial sector. Fin-R1 is built using a two-stage architecture, leveraging a financial reasoning dataset distilled and processed based on DeepSeek-R1. Through supervised fine-tuning (SFT) and reinforcement learning (RL) training, it demonstrates performance close to DeepSeek-R1 with a parameter size of 7 billion across a range of financial reasoning tasks. It achieves the state-of-the-art (SOTA) in the FinQA and ConvFinQA tasks between those LLMs in our evaluation, surpassing larger models in other tasks as well. Fin-R1 showcases strong reasoning and decision-making capabilities, providing solutions to various problems encountered in the financial domain. Our code is available at https://github.com/SUFE-AIFLM-Lab/Fin-R1. |
2025-03-21T00:00:00 | 2503.16212 | MathFusion: Enhancing Mathematic Problem-solving of LLM through Instruction Fusion | [
"Qizhi Pei",
"Lijun Wu",
"Zhuoshi Pan",
"Yu Li",
"Honglin Lin",
"Chenlin Ming",
"Xin Gao",
"Conghui He",
"Rui Yan"
] | https://github.com/QizhiPei/mathfusion | Large Language Models (LLMs) have shown impressive progress in mathematical reasoning. While data augmentation is promising to enhance mathematical problem-solving ability, current approaches are predominantly limited to instance-level modifications-such as rephrasing or generating syntactic variations-which fail to capture and leverage the intrinsic relational structures inherent in mathematical knowledge. Inspired by human learning processes, where mathematical proficiency develops through systematic exposure to interconnected concepts, we introduce MathFusion, a novel framework that enhances mathematical reasoning through cross-problem instruction synthesis. MathFusion implements this through three fusion strategies: (1) sequential fusion, which chains related problems to model solution dependencies; (2) parallel fusion, which combines analogous problems to reinforce conceptual understanding; and (3) conditional fusion, which creates context-aware selective problems to enhance reasoning flexibility. By applying these strategies, we generate a new dataset, MathFusionQA, followed by fine-tuning models (DeepSeekMath-7B, Mistral-7B, Llama3-8B) on it. Experimental results demonstrate that MathFusion achieves substantial improvements in mathematical reasoning while maintaining high data efficiency, boosting performance by 18.0 points in accuracy across diverse benchmarks while requiring only 45K additional synthetic instructions, representing a substantial improvement over traditional single-instruction approaches. Our datasets, models, and code are publicly available at https://github.com/QizhiPei/mathfusion. |
2025-03-21T00:00:00 | 2503.16031 | Deceptive Humor: A Synthetic Multilingual Benchmark Dataset for Bridging Fabricated Claims with Humorous Content | [
"Sai Kartheek Reddy Kasu",
"Shankar Biradar",
"Sunil Saumya"
] | This paper presents the Deceptive Humor Dataset (DHD), a novel resource for studying humor derived from fabricated claims and misinformation. In an era of rampant misinformation, understanding how humor intertwines with deception is essential. DHD consists of humor-infused comments generated from false narratives, incorporating fabricated claims and manipulated information using the ChatGPT-4o model. Each instance is labeled with a Satire Level, ranging from 1 for subtle satire to 3 for high-level satire and classified into five distinct Humor Categories: Dark Humor, Irony, Social Commentary, Wordplay, and Absurdity. The dataset spans multiple languages including English, Telugu, Hindi, Kannada, Tamil, and their code-mixed variants (Te-En, Hi-En, Ka-En, Ta-En), making it a valuable multilingual benchmark. By introducing DHD, we establish a structured foundation for analyzing humor in deceptive contexts, paving the way for a new research direction that explores how humor not only interacts with misinformation but also influences its perception and spread. We establish strong baselines for the proposed dataset, providing a foundation for future research to benchmark and advance deceptive humor detection models. |
|
2025-03-21T00:00:00 | 2503.15558 | Cosmos-Reason1: From Physical Common Sense To Embodied Reasoning | [
"NVIDIA",
"Alisson Azzolini",
"Hannah Brandon",
"Prithvijit Chattopadhyay",
"Huayu Chen",
"Jinju Chu",
"Yin Cui",
"Jenna Diamond",
"Yifan Ding",
"Francesco Ferroni",
"Rama Govindaraju",
"Jinwei Gu",
"Siddharth Gururani",
"Imad El Hanafi",
"Zekun Hao",
"Jacob Huffman",
"Jingyi Jin",
"Brendan Johnson",
"Rizwan Khan",
"George Kurian",
"Elena Lantz",
"Nayeon Lee",
"Zhaoshuo Li",
"Xuan Li",
"Tsung-Yi Lin",
"Yen-Chen Lin",
"Ming-Yu Liu",
"Andrew Mathau",
"Yun Ni",
"Lindsey Pavao",
"Wei Ping",
"David W. Romero",
"Misha Smelyanskiy",
"Shuran Song",
"Lyne Tchapmi",
"Andrew Z. Wang",
"Boxin Wang",
"Haoxiang Wang",
"Fangyin Wei",
"Jiashu Xu",
"Yao Xu",
"Xiaodong Yang",
"Zhuolin Yang",
"Xiaohui Zeng",
"Zhe Zhang"
] | https://github.com/nvidia-cosmos/cosmos-reason1 | Physical AI systems need to perceive, understand, and perform complex actions in the physical world. In this paper, we present the Cosmos-Reason1 models that can understand the physical world and generate appropriate embodied decisions (e.g., next step action) in natural language through long chain-of-thought reasoning processes. We begin by defining key capabilities for Physical AI reasoning, with a focus on physical common sense and embodied reasoning. To represent physical common sense, we use a hierarchical ontology that captures fundamental knowledge about space, time, and physics. For embodied reasoning, we rely on a two-dimensional ontology that generalizes across different physical embodiments. Building on these capabilities, we develop two multimodal large language models, Cosmos-Reason1-8B and Cosmos-Reason1-56B. We curate data and train our models in four stages: vision pre-training, general supervised fine-tuning (SFT), Physical AI SFT, and Physical AI reinforcement learning (RL) as the post-training. To evaluate our models, we build comprehensive benchmarks for physical common sense and embodied reasoning according to our ontologies. Evaluation results show that Physical AI SFT and reinforcement learning bring significant improvements. To facilitate the development of Physical AI, we will make our code and pre-trained models available under the NVIDIA Open Model License at https://github.com/nvidia-cosmos/cosmos-reason1. |
2025-03-21T00:00:00 | 2503.16413 | M3: 3D-Spatial MultiModal Memory | [
"Xueyan Zou",
"Yuchen Song",
"Ri-Zhao Qiu",
"Xuanbin Peng",
"Jianglong Ye",
"Sifei Liu",
"Xiaolong Wang"
] | We present 3D Spatial MultiModal Memory (M3), a multimodal memory system designed to retain information about medium-sized static scenes through video sources for visual perception. By integrating 3D Gaussian Splatting techniques with foundation models, M3 builds a multimodal memory capable of rendering feature representations across granularities, encompassing a wide range of knowledge. In our exploration, we identify two key challenges in previous works on feature splatting: (1) computational constraints in storing high-dimensional features for each Gaussian primitive, and (2) misalignment or information loss between distilled features and foundation model features. To address these challenges, we propose M3 with key components of principal scene components and Gaussian memory attention, enabling efficient training and inference. To validate M3, we conduct comprehensive quantitative evaluations of feature similarity and downstream tasks, as well as qualitative visualizations to highlight the pixel trace of Gaussian memory attention. Our approach encompasses a diverse range of foundation models, including vision-language models (VLMs), perception models, and large multimodal and language models (LMMs/LLMs). Furthermore, to demonstrate real-world applicability, we deploy M3's feature field in indoor scenes on a quadruped robot. Notably, we claim that M3 is the first work to address the core compression challenges in 3D feature distillation. |
|
2025-03-21T00:00:00 | 2503.16419 | Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models | [
"Yang Sui",
"Yu-Neng Chuang",
"Guanchu Wang",
"Jiamu Zhang",
"Tianyi Zhang",
"Jiayi Yuan",
"Hongyi Liu",
"Andrew Wen",
"Shaochen",
"Zhong",
"Hanjie Chen",
"Xia Hu"
] | Large Language Models (LLMs) have demonstrated remarkable capabilities in complex tasks. Recent advancements in Large Reasoning Models (LRMs), such as OpenAI o1 and DeepSeek-R1, have further improved performance in System-2 reasoning domains like mathematics and programming by harnessing supervised fine-tuning (SFT) and reinforcement learning (RL) techniques to enhance the Chain-of-Thought (CoT) reasoning. However, while longer CoT reasoning sequences improve performance, they also introduce significant computational overhead due to verbose and redundant outputs, known as the "overthinking phenomenon". In this paper, we provide the first structured survey to systematically investigate and explore the current progress toward achieving efficient reasoning in LLMs. Overall, relying on the inherent mechanism of LLMs, we categorize existing works into several key directions: (1) model-based efficient reasoning, which considers optimizing full-length reasoning models into more concise reasoning models or directly training efficient reasoning models; (2) reasoning output-based efficient reasoning, which aims to dynamically reduce reasoning steps and length during inference; (3) input prompts-based efficient reasoning, which seeks to enhance reasoning efficiency based on input prompt properties such as difficulty or length control. Additionally, we introduce the use of efficient data for training reasoning models, explore the reasoning capabilities of small language models, and discuss evaluation methods and benchmarking. |
|
2025-03-21T00:00:00 | 2503.16422 | 1000+ FPS 4D Gaussian Splatting for Dynamic Scene Rendering | [
"Yuheng Yuan",
"Qiuhong Shen",
"Xingyi Yang",
"Xinchao Wang"
] | 4D Gaussian Splatting (4DGS) has recently gained considerable attention as a method for reconstructing dynamic scenes. Despite achieving superior quality, 4DGS typically requires substantial storage and suffers from slow rendering speed. In this work, we delve into these issues and identify two key sources of temporal redundancy. (Q1) Short-Lifespan Gaussians: 4DGS uses a large portion of Gaussians with short temporal span to represent scene dynamics, leading to an excessive number of Gaussians. (Q2) Inactive Gaussians: When rendering, only a small subset of Gaussians contributes to each frame. Despite this, all Gaussians are processed during rasterization, resulting in redundant computation overhead. To address these redundancies, we present 4DGS-1K, which runs at over 1000 FPS on modern GPUs. For Q1, we introduce the Spatial-Temporal Variation Score, a new pruning criterion that effectively removes short-lifespan Gaussians while encouraging 4DGS to capture scene dynamics using Gaussians with longer temporal spans. For Q2, we store a mask for active Gaussians across consecutive frames, significantly reducing redundant computations in rendering. Compared to vanilla 4DGS, our method achieves a 41times reduction in storage and 9times faster rasterization speed on complex dynamic scenes, while maintaining comparable visual quality. Please see our project page at https://4DGS-1K.github.io. |
|
2025-03-21T00:00:00 | 2503.16365 | JARVIS-VLA: Post-Training Large-Scale Vision Language Models to Play Visual Games with Keyboards and Mouse | [
"Muyao Li",
"Zihao Wang",
"Kaichen He",
"Xiaojian Ma",
"Yitao Liang"
] | Recently, action-based decision-making in open-world environments has gained significant attention. Visual Language Action (VLA) models, pretrained on large-scale web datasets, have shown promise in decision-making tasks. However, previous work has primarily focused on action post-training, often neglecting enhancements to the foundational model itself. In response, we introduce a novel approach, Act from Visual Language Post-Training, which refines Visual Language Models (VLMs) through visual and linguistic guidance in a self-supervised manner. This enhancement improves the models' capabilities in world knowledge, visual recognition, and spatial grounding in open-world environments. Following the above post-training paradigms, we obtain the first VLA models in Minecraft that can follow human instructions on over 1k different atomic tasks, including crafting, smelting, cooking, mining, and killing. Our experiments demonstrate that post-training on non-trajectory tasks leads to a significant 40% improvement over the best agent baseline on a diverse set of atomic tasks. Furthermore, we demonstrate that our approach surpasses traditional imitation learning-based policies in Minecraft, achieving state-of-the-art performance. We have open-sourced the code, models, and datasets to foster further research. The project page can be found in https://craftjarvis.github.io/JarvisVLA. |
|
2025-03-21T00:00:00 | 2503.16428 | XAttention: Block Sparse Attention with Antidiagonal Scoring | [
"Ruyi Xu",
"Guangxuan Xiao",
"Haofeng Huang",
"Junxian Guo",
"Song Han"
] | https://github.com/mit-han-lab/x-attention | Long-Context Transformer Models (LCTMs) are vital for real-world applications but suffer high computational costs due to attention's quadratic complexity. Block-sparse attention mitigates this by focusing computation on critical regions, yet existing methods struggle with balancing accuracy and efficiency due to costly block importance measurements. In this paper, we introduce XAttention, a plug-and-play framework that dramatically accelerates long-context inference in Transformers models using sparse attention. XAttention's key innovation is the insight that the sum of antidiagonal values (i.e., from the lower-left to upper-right) in the attention matrix provides a powerful proxy for block importance. This allows for precise identification and pruning of non-essential blocks, resulting in high sparsity and dramatically accelerated inference. Across comprehensive evaluations on demanding long-context benchmarks-including RULER and LongBench for language, VideoMME for video understanding, and VBench for video generation. XAttention achieves accuracy comparable to full attention while delivering substantial computational gains. We demonstrate up to 13.5x acceleration in attention computation. These results underscore XAttention's ability to unlock the practical potential of block sparse attention, paving the way for scalable and efficient deployment of LCTMs in real-world applications. Code is available at https://github.com/mit-han-lab/x-attention. |
2025-03-21T00:00:00 | 2503.16421 | MagicMotion: Controllable Video Generation with Dense-to-Sparse Trajectory Guidance | [
"Quanhao Li",
"Zhen Xing",
"Rui Wang",
"Hui Zhang",
"Qi Dai",
"Zuxuan Wu"
] | Recent advances in video generation have led to remarkable improvements in visual quality and temporal coherence. Upon this, trajectory-controllable video generation has emerged to enable precise object motion control through explicitly defined spatial paths. However, existing methods struggle with complex object movements and multi-object motion control, resulting in imprecise trajectory adherence, poor object consistency, and compromised visual quality. Furthermore, these methods only support trajectory control in a single format, limiting their applicability in diverse scenarios. Additionally, there is no publicly available dataset or benchmark specifically tailored for trajectory-controllable video generation, hindering robust training and systematic evaluation. To address these challenges, we introduce MagicMotion, a novel image-to-video generation framework that enables trajectory control through three levels of conditions from dense to sparse: masks, bounding boxes, and sparse boxes. Given an input image and trajectories, MagicMotion seamlessly animates objects along defined trajectories while maintaining object consistency and visual quality. Furthermore, we present MagicData, a large-scale trajectory-controlled video dataset, along with an automated pipeline for annotation and filtering. We also introduce MagicBench, a comprehensive benchmark that assesses both video quality and trajectory control accuracy across different numbers of objects. Extensive experiments demonstrate that MagicMotion outperforms previous methods across various metrics. Our project page are publicly available at https://quanhaol.github.io/magicmotion-site. |
|
2025-03-21T00:00:00 | 2503.16302 | Unleashing Vecset Diffusion Model for Fast Shape Generation | [
"Zeqiang Lai",
"Yunfei Zhao",
"Zibo Zhao",
"Haolin Liu",
"Fuyun Wang",
"Huiwen Shi",
"Xianghui Yang",
"Qinxiang Lin",
"Jinwei Huang",
"Yuhong Liu",
"Jie Jiang",
"Chunchao Guo",
"Xiangyu Yue"
] | https://github.com/Tencent/FlashVDM | 3D shape generation has greatly flourished through the development of so-called "native" 3D diffusion, particularly through the Vecset Diffusion Model (VDM). While recent advancements have shown promising results in generating high-resolution 3D shapes, VDM still struggles with high-speed generation. Challenges exist because of difficulties not only in accelerating diffusion sampling but also VAE decoding in VDM, areas under-explored in previous works. To address these challenges, we present FlashVDM, a systematic framework for accelerating both VAE and DiT in VDM. For DiT, FlashVDM enables flexible diffusion sampling with as few as 5 inference steps and comparable quality, which is made possible by stabilizing consistency distillation with our newly introduced Progressive Flow Distillation. For VAE, we introduce a lightning vecset decoder equipped with Adaptive KV Selection, Hierarchical Volume Decoding, and Efficient Network Design. By exploiting the locality of the vecset and the sparsity of shape surface in the volume, our decoder drastically lowers FLOPs, minimizing the overall decoding overhead. We apply FlashVDM to Hunyuan3D-2 to obtain Hunyuan3D-2 Turbo. Through systematic evaluation, we show that our model significantly outperforms existing fast 3D generation methods, achieving comparable performance to the state-of-the-art while reducing inference time by over 45x for reconstruction and 32x for generation. Code and models are available at https://github.com/Tencent/FlashVDM. |
2025-03-21T00:00:00 | 2503.16322 | Ultra-Resolution Adaptation with Ease | [
"Ruonan Yu",
"Songhua Liu",
"Zhenxiong Tan",
"Xinchao Wang"
] | https://github.com/Huage001/URAE{here | Text-to-image diffusion models have achieved remarkable progress in recent years. However, training models for high-resolution image generation remains challenging, particularly when training data and computational resources are limited. In this paper, we explore this practical problem from two key perspectives: data and parameter efficiency, and propose a set of key guidelines for ultra-resolution adaptation termed URAE. For data efficiency, we theoretically and empirically demonstrate that synthetic data generated by some teacher models can significantly promote training convergence. For parameter efficiency, we find that tuning minor components of the weight matrices outperforms widely-used low-rank adapters when synthetic data are unavailable, offering substantial performance gains while maintaining efficiency. Additionally, for models leveraging guidance distillation, such as FLUX, we show that disabling classifier-free guidance, i.e., setting the guidance scale to 1 during adaptation, is crucial for satisfactory performance. Extensive experiments validate that URAE achieves comparable 2K-generation performance to state-of-the-art closed-source models like FLUX1.1 [Pro] Ultra with only 3K samples and 2K iterations, while setting new benchmarks for 4K-resolution generation. Codes are available https://github.com/Huage001/URAE{here}. |
2025-03-21T00:00:00 | 2503.16418 | InfiniteYou: Flexible Photo Recrafting While Preserving Your Identity | [
"Liming Jiang",
"Qing Yan",
"Yumin Jia",
"Zichuan Liu",
"Hao Kang",
"Xin Lu"
] | Achieving flexible and high-fidelity identity-preserved image generation remains formidable, particularly with advanced Diffusion Transformers (DiTs) like FLUX. We introduce InfiniteYou (InfU), one of the earliest robust frameworks leveraging DiTs for this task. InfU addresses significant issues of existing methods, such as insufficient identity similarity, poor text-image alignment, and low generation quality and aesthetics. Central to InfU is InfuseNet, a component that injects identity features into the DiT base model via residual connections, enhancing identity similarity while maintaining generation capabilities. A multi-stage training strategy, including pretraining and supervised fine-tuning (SFT) with synthetic single-person-multiple-sample (SPMS) data, further improves text-image alignment, ameliorates image quality, and alleviates face copy-pasting. Extensive experiments demonstrate that InfU achieves state-of-the-art performance, surpassing existing baselines. In addition, the plug-and-play design of InfU ensures compatibility with various existing methods, offering a valuable contribution to the broader community. |
|
2025-03-21T00:00:00 | 2503.12689 | MagicID: Hybrid Preference Optimization for ID-Consistent and Dynamic-Preserved Video Customization | [
"Hengjia Li",
"Lifan Jiang",
"Xi Xiao",
"Tianyang Wang",
"Hongwei Yi",
"Boxi Wu",
"Deng Cai"
] | Video identity customization seeks to produce high-fidelity videos that maintain consistent identity and exhibit significant dynamics based on users' reference images. However, existing approaches face two key challenges: identity degradation over extended video length and reduced dynamics during training, primarily due to their reliance on traditional self-reconstruction training with static images. To address these issues, we introduce MagicID, a novel framework designed to directly promote the generation of identity-consistent and dynamically rich videos tailored to user preferences. Specifically, we propose constructing pairwise preference video data with explicit identity and dynamic rewards for preference learning, instead of sticking to the traditional self-reconstruction. To address the constraints of customized preference data, we introduce a hybrid sampling strategy. This approach first prioritizes identity preservation by leveraging static videos derived from reference images, then enhances dynamic motion quality in the generated videos using a Frontier-based sampling method. By utilizing these hybrid preference pairs, we optimize the model to align with the reward differences between pairs of customized preferences. Extensive experiments show that MagicID successfully achieves consistent identity and natural dynamics, surpassing existing methods across various metrics. |
|
2025-03-21T00:00:00 | 2503.16188 | CLS-RL: Image Classification with Rule-Based Reinforcement Learning | [
"Ming Li",
"Shitian Zhao",
"Jike Zhong",
"Yuxiang Lai",
"Kaipeng Zhang"
] | Classification is a core task in machine learning. Recent research has shown that although Multimodal Large Language Models (MLLMs) are initially poor at image classification, fine-tuning them with an adequate amount of data can significantly enhance their performance, making them comparable to SOTA classification models. However, acquiring large-scale labeled data is expensive. In this paper, we explore few-shot MLLM classification fine-tuning. We found that SFT can cause severe overfitting issues and may even degrade performance over the zero-shot approach. To address this challenge, inspired by the recent successes in rule-based reinforcement learning, we propose CLS-RL, which uses verifiable signals as reward to fine-tune MLLMs. We discovered that CLS-RL outperforms SFT in most datasets and has a much higher average accuracy on both base-to-new and few-shot learning setting. Moreover, we observed a free-lunch phenomenon for CLS-RL; when models are fine-tuned on a particular dataset, their performance on other distinct datasets may also improve over zero-shot models, even if those datasets differ in distribution and class names. This suggests that RL-based methods effectively teach models the fundamentals of classification. Lastly, inspired by recent works in inference time thinking, we re-examine the `thinking process' during fine-tuning, a critical aspect of RL-based methods, in the context of visual classification. We question whether such tasks require extensive thinking process during fine-tuning, proposing that this may actually detract from performance. Based on this premise, we introduce the No-Thinking-CLS-RL method, which minimizes thinking processes during training by setting an equality accuracy reward. Our findings indicate that, with much less fine-tuning time, No-Thinking-CLS-RL method achieves superior in-domain performance and generalization capabilities than CLS-RL. |
|
2025-03-21T00:00:00 | 2503.16194 | Improving Autoregressive Image Generation through Coarse-to-Fine Token Prediction | [
"Ziyao Guo",
"Kaipeng Zhang",
"Michael Qizhe Shieh"
] | Autoregressive models have shown remarkable success in image generation by adapting sequential prediction techniques from language modeling. However, applying these approaches to images requires discretizing continuous pixel data through vector quantization methods like VQ-VAE. To alleviate the quantization errors that existed in VQ-VAE, recent works tend to use larger codebooks. However, this will accordingly expand vocabulary size, complicating the autoregressive modeling task. This paper aims to find a way to enjoy the benefits of large codebooks without making autoregressive modeling more difficult. Through empirical investigation, we discover that tokens with similar codeword representations produce similar effects on the final generated image, revealing significant redundancy in large codebooks. Based on this insight, we propose to predict tokens from coarse to fine (CTF), realized by assigning the same coarse label for similar tokens. Our framework consists of two stages: (1) an autoregressive model that sequentially predicts coarse labels for each token in the sequence, and (2) an auxiliary model that simultaneously predicts fine-grained labels for all tokens conditioned on their coarse labels. Experiments on ImageNet demonstrate our method's superior performance, achieving an average improvement of 59 points in Inception Score compared to baselines. Notably, despite adding an inference step, our approach achieves faster sampling speeds. |
|
2025-03-21T00:00:00 | 2503.13834 | See-Saw Modality Balance: See Gradient, and Sew Impaired Vision-Language Balance to Mitigate Dominant Modality Bias | [
"JuneHyoung Kwon",
"MiHyeon Kim",
"Eunju Lee",
"Juhwan Choi",
"YoungBin Kim"
] | Vision-language (VL) models have demonstrated strong performance across various tasks. However, these models often rely on a specific modality for predictions, leading to "dominant modality bias.'' This bias significantly hurts performance, especially when one modality is impaired. In this study, we analyze model behavior under dominant modality bias and theoretically show that unaligned gradients or differences in gradient magnitudes prevent balanced convergence of the loss. Based on these findings, we propose a novel framework, BalGrad to mitigate dominant modality bias. Our approach includes inter-modality gradient reweighting, adjusting the gradient of KL divergence based on each modality's contribution, and inter-task gradient projection to align task directions in a non-conflicting manner. Experiments on UPMC Food-101, Hateful Memes, and MM-IMDb datasets confirm that BalGrad effectively alleviates over-reliance on specific modalities when making predictions. |
|
2025-03-21T00:00:00 | 2503.16057 | Expert Race: A Flexible Routing Strategy for Scaling Diffusion Transformer with Mixture of Experts | [
"Yike Yuan",
"Ziyu Wang",
"Zihao Huang",
"Defa Zhu",
"Xun Zhou",
"Jingyi Yu",
"Qiyang Min"
] | Diffusion models have emerged as mainstream framework in visual generation. Building upon this success, the integration of Mixture of Experts (MoE) methods has shown promise in enhancing model scalability and performance. In this paper, we introduce Race-DiT, a novel MoE model for diffusion transformers with a flexible routing strategy, Expert Race. By allowing tokens and experts to compete together and select the top candidates, the model learns to dynamically assign experts to critical tokens. Additionally, we propose per-layer regularization to address challenges in shallow layer learning, and router similarity loss to prevent mode collapse, ensuring better expert utilization. Extensive experiments on ImageNet validate the effectiveness of our approach, showcasing significant performance gains while promising scaling properties. |
|
2025-03-21T00:00:00 | 2503.15851 | Zero-1-to-A: Zero-Shot One Image to Animatable Head Avatars Using Video Diffusion | [
"Zhou Zhenglin",
"Ma Fan",
"Fan Hehe",
"Chua Tat-Seng"
] | https://github.com/ZhenglinZhou/Zero-1-to-A | Animatable head avatar generation typically requires extensive data for training. To reduce the data requirements, a natural solution is to leverage existing data-free static avatar generation methods, such as pre-trained diffusion models with score distillation sampling (SDS), which align avatars with pseudo ground-truth outputs from the diffusion model. However, directly distilling 4D avatars from video diffusion often leads to over-smooth results due to spatial and temporal inconsistencies in the generated video. To address this issue, we propose Zero-1-to-A, a robust method that synthesizes a spatial and temporal consistency dataset for 4D avatar reconstruction using the video diffusion model. Specifically, Zero-1-to-A iteratively constructs video datasets and optimizes animatable avatars in a progressive manner, ensuring that avatar quality increases smoothly and consistently throughout the learning process. This progressive learning involves two stages: (1) Spatial Consistency Learning fixes expressions and learns from front-to-side views, and (2) Temporal Consistency Learning fixes views and learns from relaxed to exaggerated expressions, generating 4D avatars in a simple-to-complex manner. Extensive experiments demonstrate that Zero-1-to-A improves fidelity, animation quality, and rendering speed compared to existing diffusion-based methods, providing a solution for lifelike avatar creation. Code is publicly available at: https://github.com/ZhenglinZhou/Zero-1-to-A. |
2025-03-21T00:00:00 | 2503.15451 | MotionStreamer: Streaming Motion Generation via Diffusion-based Autoregressive Model in Causal Latent Space | [
"Lixing Xiao",
"Shunlin Lu",
"Huaijin Pi",
"Ke Fan",
"Liang Pan",
"Yueer Zhou",
"Ziyong Feng",
"Xiaowei Zhou",
"Sida Peng",
"Jingbo Wang"
] | This paper addresses the challenge of text-conditioned streaming motion generation, which requires us to predict the next-step human pose based on variable-length historical motions and incoming texts. Existing methods struggle to achieve streaming motion generation, e.g., diffusion models are constrained by pre-defined motion lengths, while GPT-based methods suffer from delayed response and error accumulation problem due to discretized non-causal tokenization. To solve these problems, we propose MotionStreamer, a novel framework that incorporates a continuous causal latent space into a probabilistic autoregressive model. The continuous latents mitigate information loss caused by discretization and effectively reduce error accumulation during long-term autoregressive generation. In addition, by establishing temporal causal dependencies between current and historical motion latents, our model fully utilizes the available information to achieve accurate online motion decoding. Experiments show that our method outperforms existing approaches while offering more applications, including multi-round generation, long-term generation, and dynamic motion composition. Project Page: https://zju3dv.github.io/MotionStreamer/ |
|
2025-03-21T00:00:00 | 2503.16055 | SALT: Singular Value Adaptation with Low-Rank Transformation | [
"Abdelrahman Elsayed",
"Sarim Hashmi",
"Mohammed Elseiagy",
"Hu Wang",
"Mohammad Yaqub",
"Ibrahim Almakky"
] | https://github.com/BioMedIA-MBZUAI/SALT | The complex nature of medical image segmentation calls for models that are specifically designed to capture detailed, domain-specific features. Large foundation models offer considerable flexibility, yet the cost of fine-tuning these models remains a significant barrier. Parameter-Efficient Fine-Tuning (PEFT) methods, such as Low-Rank Adaptation (LoRA), efficiently update model weights with low-rank matrices but may suffer from underfitting when the chosen rank is insufficient to capture domain-specific nuances. Conversely, full-rank Singular Value Decomposition (SVD) based methods provide comprehensive updates by modifying all singular values, yet they often lack flexibility and exhibit variable performance across datasets. We propose SALT (Singular Value Adaptation with Low-Rank Transformation), a method that selectively adapts the most influential singular values using trainable scale and shift parameters while complementing this with a low-rank update for the remaining subspace. This hybrid approach harnesses the advantages of both LoRA and SVD, enabling effective adaptation without relying on increasing model size or depth. Evaluated on 5 challenging medical datasets, ranging from as few as 20 samples to 1000, SALT outperforms state-of-the-art PEFT (LoRA and SVD) by 2% to 5% in Dice with only 3.9% trainable parameters, demonstrating robust adaptation even in low-resource settings. The code for SALT is available at: https://github.com/BioMedIA-MBZUAI/SALT |
2025-03-21T00:00:00 | 2503.10625 | LHM: Large Animatable Human Reconstruction Model from a Single Image in Seconds | [
"Lingteng Qiu",
"Xiaodong Gu",
"Peihao Li",
"Qi Zuo",
"Weichao Shen",
"Junfei Zhang",
"Kejie Qiu",
"Weihao Yuan",
"Guanying Chen",
"Zilong Dong",
"Liefeng Bo"
] | Animatable 3D human reconstruction from a single image is a challenging problem due to the ambiguity in decoupling geometry, appearance, and deformation. Recent advances in 3D human reconstruction mainly focus on static human modeling, and the reliance of using synthetic 3D scans for training limits their generalization ability. Conversely, optimization-based video methods achieve higher fidelity but demand controlled capture conditions and computationally intensive refinement processes. Motivated by the emergence of large reconstruction models for efficient static reconstruction, we propose LHM (Large Animatable Human Reconstruction Model) to infer high-fidelity avatars represented as 3D Gaussian splatting in a feed-forward pass. Our model leverages a multimodal transformer architecture to effectively encode the human body positional features and image features with attention mechanism, enabling detailed preservation of clothing geometry and texture. To further boost the face identity preservation and fine detail recovery, we propose a head feature pyramid encoding scheme to aggregate multi-scale features of the head regions. Extensive experiments demonstrate that our LHM generates plausible animatable human in seconds without post-processing for face and hands, outperforming existing methods in both reconstruction accuracy and generalization ability. |
|
2025-03-21T00:00:00 | 2503.14487 | DiffMoE: Dynamic Token Selection for Scalable Diffusion Transformers | [
"Minglei Shi",
"Ziyang Yuan",
"Haotian Yang",
"Xintao Wang",
"Mingwu Zheng",
"Xin Tao",
"Wenliang Zhao",
"Wenzhao Zheng",
"Jie Zhou",
"Jiwen Lu",
"Pengfei Wan",
"Di Zhang",
"Kun Gai"
] | Diffusion models have demonstrated remarkable success in various image generation tasks, but their performance is often limited by the uniform processing of inputs across varying conditions and noise levels. To address this limitation, we propose a novel approach that leverages the inherent heterogeneity of the diffusion process. Our method, DiffMoE, introduces a batch-level global token pool that enables experts to access global token distributions during training, promoting specialized expert behavior. To unleash the full potential of the diffusion process, DiffMoE incorporates a capacity predictor that dynamically allocates computational resources based on noise levels and sample complexity. Through comprehensive evaluation, DiffMoE achieves state-of-the-art performance among diffusion models on ImageNet benchmark, substantially outperforming both dense architectures with 3x activated parameters and existing MoE approaches while maintaining 1x activated parameters. The effectiveness of our approach extends beyond class-conditional generation to more challenging tasks such as text-to-image generation, demonstrating its broad applicability across different diffusion model applications. Project Page: https://shiml20.github.io/DiffMoE/ |
|
2025-03-21T00:00:00 | 2503.16416 | Survey on Evaluation of LLM-based Agents | [
"Asaf Yehudai",
"Lilach Eden",
"Alan Li",
"Guy Uziel",
"Yilun Zhao",
"Roy Bar-Haim",
"Arman Cohan",
"Michal Shmueli-Scheuer"
] | The emergence of LLM-based agents represents a paradigm shift in AI, enabling autonomous systems to plan, reason, use tools, and maintain memory while interacting with dynamic environments. This paper provides the first comprehensive survey of evaluation methodologies for these increasingly capable agents. We systematically analyze evaluation benchmarks and frameworks across four critical dimensions: (1) fundamental agent capabilities, including planning, tool use, self-reflection, and memory; (2) application-specific benchmarks for web, software engineering, scientific, and conversational agents; (3) benchmarks for generalist agents; and (4) frameworks for evaluating agents. Our analysis reveals emerging trends, including a shift toward more realistic, challenging evaluations with continuously updated benchmarks. We also identify critical gaps that future research must address-particularly in assessing cost-efficiency, safety, and robustness, and in developing fine-grained, and scalable evaluation methods. This survey maps the rapidly evolving landscape of agent evaluation, reveals the emerging trends in the field, identifies current limitations, and proposes directions for future research. |
|
2025-03-21T00:00:00 | 2503.13657 | Why Do Multi-Agent LLM Systems Fail? | [
"Mert Cemri",
"Melissa Z. Pan",
"Shuyi Yang",
"Lakshya A. Agrawal",
"Bhavya Chopra",
"Rishabh Tiwari",
"Kurt Keutzer",
"Aditya Parameswaran",
"Dan Klein",
"Kannan Ramchandran",
"Matei Zaharia",
"Joseph E. Gonzalez",
"Ion Stoica"
] | Despite growing enthusiasm for Multi-Agent Systems (MAS), where multiple LLM agents collaborate to accomplish tasks, their performance gains across popular benchmarks remain minimal compared to single-agent frameworks. This gap highlights the need to analyze the challenges hindering MAS effectiveness. In this paper, we present the first comprehensive study of MAS challenges. We analyze five popular MAS frameworks across over 150 tasks, involving six expert human annotators. We identify 14 unique failure modes and propose a comprehensive taxonomy applicable to various MAS frameworks. This taxonomy emerges iteratively from agreements among three expert annotators per study, achieving a Cohen's Kappa score of 0.88. These fine-grained failure modes are organized into 3 categories, (i) specification and system design failures, (ii) inter-agent misalignment, and (iii) task verification and termination. To support scalable evaluation, we integrate MASFT with LLM-as-a-Judge. We also explore if identified failures could be easily prevented by proposing two interventions: improved specification of agent roles and enhanced orchestration strategies. Our findings reveal that identified failures require more complex solutions, highlighting a clear roadmap for future research. We open-source our dataset and LLM annotator. |
|
2025-03-21T00:00:00 | 2503.16375 | NuiScene: Exploring Efficient Generation of Unbounded Outdoor Scenes | [
"Han-Hung Lee",
"Qinghong Han",
"Angel X. Chang"
] | In this paper, we explore the task of generating expansive outdoor scenes, ranging from castles to high-rises. Unlike indoor scene generation, which has been a primary focus of prior work, outdoor scene generation presents unique challenges, including wide variations in scene heights and the need for a method capable of rapidly producing large landscapes. To address this, we propose an efficient approach that encodes scene chunks as uniform vector sets, offering better compression and performance than the spatially structured latents used in prior methods. Furthermore, we train an explicit outpainting model for unbounded generation, which improves coherence compared to prior resampling-based inpainting schemes while also speeding up generation by eliminating extra diffusion steps. To facilitate this task, we curate NuiScene43, a small but high-quality set of scenes, preprocessed for joint training. Notably, when trained on scenes of varying styles, our model can blend different environments, such as rural houses and city skyscrapers, within the same scene, highlighting the potential of our curation process to leverage heterogeneous scenes for joint training. |
|
2025-03-21T00:00:00 | 2503.16219 | Reinforcement Learning for Reasoning in Small LLMs: What Works and What Doesn't | [
"Quy-Anh Dang",
"Chris Ngo"
] | https://github.com/knoveleng/open-rs | Enhancing the reasoning capabilities of large language models (LLMs) typically relies on massive computational resources and extensive datasets, limiting accessibility for resource-constrained settings. Our study investigates the potential of reinforcement learning (RL) to improve reasoning in small LLMs, focusing on a 1.5-billion-parameter model, DeepSeek-R1-Distill-Qwen-1.5B, under strict constraints: training on 4 NVIDIA A40 GPUs (48 GB VRAM each) within 24 hours. Adapting the Group Relative Policy Optimization (GRPO) algorithm and curating a compact, high-quality mathematical reasoning dataset, we conducted three experiments to explore model behavior and performance. Our results demonstrate rapid reasoning gains - e.g., AMC23 accuracy rising from 63% to 80% and AIME24 reaching 46.7%, surpassing o1-preview - using only 7,000 samples and a $42 training cost, compared to thousands of dollars for baseline models. However, challenges such as optimization instability and length constraints emerged with prolonged training. These findings highlight the efficacy of RL-based fine-tuning for small LLMs, offering a cost-effective alternative to large-scale approaches. We release our code and datasets as open-source resources, providing insights into trade-offs and laying a foundation for scalable, reasoning-capable LLMs in resource-limited environments. All are available at https://github.com/knoveleng/open-rs. |
2025-03-21T00:00:00 | 2503.13358 | One-Step Residual Shifting Diffusion for Image Super-Resolution via Distillation | [
"Daniil Selikhanovych",
"David Li",
"Aleksei Leonov",
"Nikita Gushchin",
"Sergei Kushneriuk",
"Alexander Filippov",
"Evgeny Burnaev",
"Iaroslav Koshelev",
"Alexander Korotin"
] | Diffusion models for super-resolution (SR) produce high-quality visual results but require expensive computational costs. Despite the development of several methods to accelerate diffusion-based SR models, some (e.g., SinSR) fail to produce realistic perceptual details, while others (e.g., OSEDiff) may hallucinate non-existent structures. To overcome these issues, we present RSD, a new distillation method for ResShift, one of the top diffusion-based SR models. Our method is based on training the student network to produce such images that a new fake ResShift model trained on them will coincide with the teacher model. RSD achieves single-step restoration and outperforms the teacher by a large margin. We show that our distillation method can surpass the other distillation-based method for ResShift - SinSR - making it on par with state-of-the-art diffusion-based SR distillation methods. Compared to SR methods based on pre-trained text-to-image models, RSD produces competitive perceptual quality, provides images with better alignment to degraded input images, and requires fewer parameters and GPU memory. We provide experimental results on various real-world and synthetic datasets, including RealSR, RealSet65, DRealSR, ImageNet, and DIV2K. |
|
2025-03-21T00:00:00 | 2503.16257 | Plug-and-Play 1.x-Bit KV Cache Quantization for Video Large Language Models | [
"Keda Tao",
"Haoxuan You",
"Yang Sui",
"Can Qin",
"Huan Wang"
] | Video large language models (VideoLLMs) have demonstrated the capability to process longer video inputs and enable complex reasoning and analysis. However, due to the thousands of visual tokens from the video frames, key-value (KV) cache can significantly increase memory requirements, becoming a bottleneck for inference speed and memory usage. KV cache quantization is a widely used approach to address this problem. In this paper, we find that 2-bit KV quantization of VideoLLMs can hardly hurt the model performance, while the limit of KV cache quantization in even lower bits has not been investigated. To bridge this gap, we introduce VidKV, a plug-and-play KV cache quantization method to compress the KV cache to lower than 2 bits. Specifically, (1) for key, we propose a mixed-precision quantization strategy in the channel dimension, where we perform 2-bit quantization for anomalous channels and 1-bit quantization combined with FFT for normal channels; (2) for value, we implement 1.58-bit quantization while selectively filtering semantically salient visual tokens for targeted preservation, for a better trade-off between precision and model performance. Importantly, our findings suggest that the value cache of VideoLLMs should be quantized in a per-channel fashion instead of the per-token fashion proposed by prior KV cache quantization works for LLMs. Empirically, extensive results with LLaVA-OV-7B and Qwen2.5-VL-7B on six benchmarks show that VidKV effectively compresses the KV cache to 1.5-bit and 1.58-bit precision with almost no performance drop compared to the FP16 counterparts. |
|
2025-03-21T00:00:00 | 2503.16397 | Scale-wise Distillation of Diffusion Models | [
"Nikita Starodubcev",
"Denis Kuznedelev",
"Artem Babenko",
"Dmitry Baranchuk"
] | We present SwD, a scale-wise distillation framework for diffusion models (DMs), which effectively employs next-scale prediction ideas for diffusion-based few-step generators. In more detail, SwD is inspired by the recent insights relating diffusion processes to the implicit spectral autoregression. We suppose that DMs can initiate generation at lower data resolutions and gradually upscale the samples at each denoising step without loss in performance while significantly reducing computational costs. SwD naturally integrates this idea into existing diffusion distillation methods based on distribution matching. Also, we enrich the family of distribution matching approaches by introducing a novel patch loss enforcing finer-grained similarity to the target distribution. When applied to state-of-the-art text-to-image diffusion models, SwD approaches the inference times of two full resolution steps and significantly outperforms the counterparts under the same computation budget, as evidenced by automated metrics and human preference studies. |
|
2025-03-21T00:00:00 | 2503.15855 | VideoRFSplat: Direct Scene-Level Text-to-3D Gaussian Splatting Generation with Flexible Pose and Multi-View Joint Modeling | [
"Hyojun Go",
"Byeongjun Park",
"Hyelin Nam",
"Byung-Hoon Kim",
"Hyungjin Chung",
"Changick Kim"
] | We propose VideoRFSplat, a direct text-to-3D model leveraging a video generation model to generate realistic 3D Gaussian Splatting (3DGS) for unbounded real-world scenes. To generate diverse camera poses and unbounded spatial extent of real-world scenes, while ensuring generalization to arbitrary text prompts, previous methods fine-tune 2D generative models to jointly model camera poses and multi-view images. However, these methods suffer from instability when extending 2D generative models to joint modeling due to the modality gap, which necessitates additional models to stabilize training and inference. In this work, we propose an architecture and a sampling strategy to jointly model multi-view images and camera poses when fine-tuning a video generation model. Our core idea is a dual-stream architecture that attaches a dedicated pose generation model alongside a pre-trained video generation model via communication blocks, generating multi-view images and camera poses through separate streams. This design reduces interference between the pose and image modalities. Additionally, we propose an asynchronous sampling strategy that denoises camera poses faster than multi-view images, allowing rapidly denoised poses to condition multi-view generation, reducing mutual ambiguity and enhancing cross-modal consistency. Trained on multiple large-scale real-world datasets (RealEstate10K, MVImgNet, DL3DV-10K, ACID), VideoRFSplat outperforms existing text-to-3D direct generation methods that heavily depend on post-hoc refinement via score distillation sampling, achieving superior results without such refinement. |
|
2025-03-21T00:00:00 | 2503.14237 | Make Your Training Flexible: Towards Deployment-Efficient Video Models | [
"Chenting Wang",
"Kunchang Li",
"Tianxiang Jiang",
"Xiangyu Zeng",
"Yi Wang",
"Limin Wang"
] | https://github.com/OpenGVLab/FluxViT | Popular video training methods mainly operate on a fixed number of tokens sampled from a predetermined spatiotemporal grid, resulting in sub-optimal accuracy-computation trade-offs due to inherent video redundancy. They also lack adaptability to varying computational budgets for downstream tasks, hindering applications of the most competitive model in real-world scenes. We thus propose a new test setting, Token Optimization, for maximized input information across budgets, which optimizes the size-limited set of input tokens through token selection from more suitably sampled videos. To this end, we propose a novel augmentation tool termed Flux. By making the sampling grid flexible and leveraging token selection, it is easily adopted in most popular video training frameworks, boosting model robustness with nearly no additional cost. We integrate Flux in large-scale video pre-training, and the resulting FluxViT establishes new state-of-the-art results across extensive tasks at standard costs. Notably, with 1/4 tokens only, it can still match the performance of previous state-of-the-art models with Token Optimization, yielding nearly 90\% savings. All models and data are available at https://github.com/OpenGVLab/FluxViT. |
2025-03-21T00:00:00 | 2503.16425 | Tokenize Image as a Set | [
"Zigang Geng",
"Mengde Xu",
"Han Hu",
"Shuyang Gu"
] | https://github.com/Gengzigang/TokenSet | This paper proposes a fundamentally new paradigm for image generation through set-based tokenization and distribution modeling. Unlike conventional methods that serialize images into fixed-position latent codes with a uniform compression ratio, we introduce an unordered token set representation to dynamically allocate coding capacity based on regional semantic complexity. This TokenSet enhances global context aggregation and improves robustness against local perturbations. To address the critical challenge of modeling discrete sets, we devise a dual transformation mechanism that bijectively converts sets into fixed-length integer sequences with summation constraints. Further, we propose Fixed-Sum Discrete Diffusion--the first framework to simultaneously handle discrete values, fixed sequence length, and summation invariance--enabling effective set distribution modeling. Experiments demonstrate our method's superiority in semantic-aware representation and generation quality. Our innovations, spanning novel representation and modeling strategies, advance visual generation beyond traditional sequential token paradigms. Our code and models are publicly available at https://github.com/Gengzigang/TokenSet. |
2025-03-21T00:00:00 | 2503.15242 | BigO(Bench) -- Can LLMs Generate Code with Controlled Time and Space Complexity? | [
"Pierre Chambon",
"Baptiste Roziere",
"Benoit Sagot",
"Gabriel Synnaeve"
] | We introduce BigO(Bench), a novel coding benchmark designed to evaluate the capabilities of generative language models in understanding and generating code with specified time and space complexities. This benchmark addresses the gap in current evaluations that often overlook the ability of models to comprehend and produce code constrained by computational complexity. BigO(Bench) includes tooling to infer the algorithmic complexity of any Python function from profiling measurements, including human- or LLM-generated solutions. BigO(Bench) also includes of set of 3,105 coding problems and 1,190,250 solutions from Code Contests annotated with inferred (synthetic) time and space complexity labels from the complexity framework, as well as corresponding runtime and memory footprint values for a large set of input sizes. We present results from evaluating multiple state-of-the-art language models on this benchmark, highlighting their strengths and weaknesses in handling complexity requirements. In particular, token-space reasoning models are unrivaled in code generation but not in complexity understanding, hinting that they may not generalize well to tasks for which no reward was given at training time. |
|
2025-03-21T00:00:00 | 2503.16091 | AIMI: Leveraging Future Knowledge and Personalization in Sparse Event Forecasting for Treatment Adherence | [
"Abdullah Mamun",
"Diane J. Cook",
"Hassan Ghasemzadeh"
] | https://github.com/ab9mamun/AIMI | Adherence to prescribed treatments is crucial for individuals with chronic conditions to avoid costly or adverse health outcomes. For certain patient groups, intensive lifestyle interventions are vital for enhancing medication adherence. Accurate forecasting of treatment adherence can open pathways to developing an on-demand intervention tool, enabling timely and personalized support. With the increasing popularity of smartphones and wearables, it is now easier than ever to develop and deploy smart activity monitoring systems. However, effective forecasting systems for treatment adherence based on wearable sensors are still not widely available. We close this gap by proposing Adherence Forecasting and Intervention with Machine Intelligence (AIMI). AIMI is a knowledge-guided adherence forecasting system that leverages smartphone sensors and previous medication history to estimate the likelihood of forgetting to take a prescribed medication. A user study was conducted with 27 participants who took daily medications to manage their cardiovascular diseases. We designed and developed CNN and LSTM-based forecasting models with various combinations of input features and found that LSTM models can forecast medication adherence with an accuracy of 0.932 and an F-1 score of 0.936. Moreover, through a series of ablation studies involving convolutional and recurrent neural network architectures, we demonstrate that leveraging known knowledge about future and personalized training enhances the accuracy of medication adherence forecasting. Code available: https://github.com/ab9mamun/AIMI. |
2025-03-21T00:00:00 | 2503.15299 | Inside-Out: Hidden Factual Knowledge in LLMs | [
"Zorik Gekhman",
"Eyal Ben David",
"Hadas Orgad",
"Eran Ofek",
"Yonatan Belinkov",
"Idan Szpector",
"Jonathan Herzig",
"Roi Reichart"
] | This work presents a framework for assessing whether large language models (LLMs) encode more factual knowledge in their parameters than what they express in their outputs. While a few studies hint at this possibility, none has clearly defined or demonstrated this phenomenon. We first propose a formal definition of knowledge, quantifying it for a given question as the fraction of correct-incorrect answer pairs where the correct one is ranked higher. This gives rise to external and internal knowledge, depending on the information used to score individual answer candidates: either the model's observable token-level probabilities or its intermediate computations. Hidden knowledge arises when internal knowledge exceeds external knowledge. We then present a case study, applying this framework to three popular open-weights LLMs in a closed-book QA setup. Our results indicate that: (1) LLMs consistently encode more factual knowledge internally than what they express externally, with an average gap of 40%. (2) Surprisingly, some knowledge is so deeply hidden that a model can internally know an answer perfectly, yet fail to generate it even once, despite large-scale repeated sampling of 1,000 answers. This reveals fundamental limitations in the generation capabilities of LLMs, which (3) puts a practical constraint on scaling test-time compute via repeated answer sampling in closed-book QA: significant performance improvements remain inaccessible because some answers are practically never sampled, yet if they were, we would be guaranteed to rank them first. |
|
2025-03-21T00:00:00 | 2503.16429 | Sonata: Self-Supervised Learning of Reliable Point Representations | [
"Xiaoyang Wu",
"Daniel DeTone",
"Duncan Frost",
"Tianwei Shen",
"Chris Xie",
"Nan Yang",
"Jakob Engel",
"Richard Newcombe",
"Hengshuang Zhao",
"Julian Straub"
] | In this paper, we question whether we have a reliable self-supervised point cloud model that can be used for diverse 3D tasks via simple linear probing, even with limited data and minimal computation. We find that existing 3D self-supervised learning approaches fall short when evaluated on representation quality through linear probing. We hypothesize that this is due to what we term the "geometric shortcut", which causes representations to collapse to low-level spatial features. This challenge is unique to 3D and arises from the sparse nature of point cloud data. We address it through two key strategies: obscuring spatial information and enhancing the reliance on input features, ultimately composing a Sonata of 140k point clouds through self-distillation. Sonata is simple and intuitive, yet its learned representations are strong and reliable: zero-shot visualizations demonstrate semantic grouping, alongside strong spatial reasoning through nearest-neighbor relationships. Sonata demonstrates exceptional parameter and data efficiency, tripling linear probing accuracy (from 21.8% to 72.5%) on ScanNet and nearly doubling performance with only 1% of the data compared to previous approaches. Full fine-tuning further advances SOTA across both 3D indoor and outdoor perception tasks. |
|
2025-03-21T00:00:00 | 2503.09949 | UVE: Are MLLMs Unified Evaluators for AI-Generated Videos? | [
"Yuanxin Liu",
"Rui Zhu",
"Shuhuai Ren",
"Jiacong Wang",
"Haoyuan Guo",
"Xu Sun",
"Lu Jiang"
] | https://github.com/bytedance/UVE | With the rapid growth of video generative models (VGMs), it is essential to develop reliable and comprehensive automatic metrics for AI-generated videos (AIGVs). Existing methods either use off-the-shelf models optimized for other tasks or rely on human assessment data to train specialized evaluators. These approaches are constrained to specific evaluation aspects and are difficult to scale with the increasing demands for finer-grained and more comprehensive evaluations. To address this issue, this work investigates the feasibility of using multimodal large language models (MLLMs) as a unified evaluator for AIGVs, leveraging their strong visual perception and language understanding capabilities. To evaluate the performance of automatic metrics in unified AIGV evaluation, we introduce a benchmark called UVE-Bench. UVE-Bench collects videos generated by state-of-the-art VGMs and provides pairwise human preference annotations across 15 evaluation aspects. Using UVE-Bench, we extensively evaluate 16 MLLMs. Our empirical results suggest that while advanced MLLMs (e.g., Qwen2VL-72B and InternVL2.5-78B) still lag behind human evaluators, they demonstrate promising ability in unified AIGV evaluation, significantly surpassing existing specialized evaluation methods. Additionally, we conduct an in-depth analysis of key design choices that impact the performance of MLLM-driven evaluators, offering valuable insights for future research on AIGV evaluation. The code is available at https://github.com/bytedance/UVE. |
2025-03-21T00:00:00 | 2503.13356 | Agents Play Thousands of 3D Video Games | [
"Zhongwen Xu",
"Xianliang Wang",
"Siyi Li",
"Tao Yu",
"Liang Wang",
"Qiang Fu",
"Wei Yang"
] | We present PORTAL, a novel framework for developing artificial intelligence agents capable of playing thousands of 3D video games through language-guided policy generation. By transforming decision-making problems into language modeling tasks, our approach leverages large language models (LLMs) to generate behavior trees represented in domain-specific language (DSL). This method eliminates the computational burden associated with traditional reinforcement learning approaches while preserving strategic depth and rapid adaptability. Our framework introduces a hybrid policy structure that combines rule-based nodes with neural network components, enabling both high-level strategic reasoning and precise low-level control. A dual-feedback mechanism incorporating quantitative game metrics and vision-language model analysis facilitates iterative policy improvement at both tactical and strategic levels. The resulting policies are instantaneously deployable, human-interpretable, and capable of generalizing across diverse gaming environments. Experimental results demonstrate PORTAL's effectiveness across thousands of first-person shooter (FPS) games, showcasing significant improvements in development efficiency, policy generalization, and behavior diversity compared to traditional approaches. PORTAL represents a significant advancement in game AI development, offering a practical solution for creating sophisticated agents that can operate across thousands of commercial video games with minimal development overhead. Experiment results on the 3D video games are best viewed on https://zhongwen.one/projects/portal . |
|
2025-03-21T00:00:00 | 2503.16420 | SynCity: Training-Free Generation of 3D Worlds | [
"Paul Engstler",
"Aleksandar Shtedritski",
"Iro Laina",
"Christian Rupprecht",
"Andrea Vedaldi"
] | We address the challenge of generating 3D worlds from textual descriptions. We propose SynCity, a training- and optimization-free approach, which leverages the geometric precision of pre-trained 3D generative models and the artistic versatility of 2D image generators to create large, high-quality 3D spaces. While most 3D generative models are object-centric and cannot generate large-scale worlds, we show how 3D and 2D generators can be combined to generate ever-expanding scenes. Through a tile-based approach, we allow fine-grained control over the layout and the appearance of scenes. The world is generated tile-by-tile, and each new tile is generated within its world-context and then fused with the scene. SynCity generates compelling and immersive scenes that are rich in detail and diversity. |
|
2025-03-21T00:00:00 | 2503.11509 | TikZero: Zero-Shot Text-Guided Graphics Program Synthesis | [
"Jonas Belouadi",
"Eddy Ilg",
"Margret Keuper",
"Hideki Tanaka",
"Masao Utiyama",
"Raj Dabre",
"Steffen Eger",
"Simone Paolo Ponzetto"
] | With the rise of generative AI, synthesizing figures from text captions becomes a compelling application. However, achieving high geometric precision and editability requires representing figures as graphics programs in languages like TikZ, and aligned training data (i.e., graphics programs with captions) remains scarce. Meanwhile, large amounts of unaligned graphics programs and captioned raster images are more readily available. We reconcile these disparate data sources by presenting TikZero, which decouples graphics program generation from text understanding by using image representations as an intermediary bridge. It enables independent training on graphics programs and captioned images and allows for zero-shot text-guided graphics program synthesis during inference. We show that our method substantially outperforms baselines that can only operate with caption-aligned graphics programs. Furthermore, when leveraging caption-aligned graphics programs as a complementary training signal, TikZero matches or exceeds the performance of much larger models, including commercial systems like GPT-4o. Our code, datasets, and select models are publicly available. |
|
2025-03-21T00:00:00 | 2503.15672 | GASP: Unifying Geometric and Semantic Self-Supervised Pre-training for Autonomous Driving | [
"William Ljungbergh",
"Adam Lilja",
"Adam Tonderski. Arvid Laveno Ling",
"Carl Lindström",
"Willem Verbeke",
"Junsheng Fu",
"Christoffer Petersson",
"Lars Hammarstrand",
"Michael Felsberg"
] | Self-supervised pre-training based on next-token prediction has enabled large language models to capture the underlying structure of text, and has led to unprecedented performance on a large array of tasks when applied at scale. Similarly, autonomous driving generates vast amounts of spatiotemporal data, alluding to the possibility of harnessing scale to learn the underlying geometric and semantic structure of the environment and its evolution over time. In this direction, we propose a geometric and semantic self-supervised pre-training method, GASP, that learns a unified representation by predicting, at any queried future point in spacetime, (1) general occupancy, capturing the evolving structure of the 3D scene; (2) ego occupancy, modeling the ego vehicle path through the environment; and (3) distilled high-level features from a vision foundation model. By modeling geometric and semantic 4D occupancy fields instead of raw sensor measurements, the model learns a structured, generalizable representation of the environment and its evolution through time. We validate GASP on multiple autonomous driving benchmarks, demonstrating significant improvements in semantic occupancy forecasting, online mapping, and ego trajectory prediction. Our results demonstrate that continuous 4D geometric and semantic occupancy prediction provides a scalable and effective pre-training paradigm for autonomous driving. For code and additional visualizations, see \href{https://research.zenseact.com/publications/gasp/. |
|
2025-03-21T00:00:00 | 2503.14201 | Why Personalizing Deep Learning-Based Code Completion Tools Matters | [
"Alessandro Giagnorio",
"Alberto Martin-Lopez",
"Gabriele Bavota"
] | Deep learning (DL)-based code completion tools have transformed software development by enabling advanced code generation. These tools leverage models trained on vast amounts of code from numerous repositories, capturing general coding patterns. However, the impact of fine-tuning these models for specific organizations or developers to boost their performance on such subjects remains unexplored. In this work, we fill this gap by presenting solid empirical evidence answering this question. More specifically, we consider 136 developers from two organizations (Apache and Spring), two model architectures (T5 and Code Llama), and three model sizes (60M, 750M, and 7B trainable parameters). T5 models (60M, 750M) were pre-trained and fine-tuned on over 2,000 open-source projects, excluding the subject organizations' data, and compared against versions fine-tuned on organization- and developer-specific datasets. For the Code Llama model (7B), we compared the performance of the already pre-trained model publicly available online with the same model fine-tuned via parameter-efficient fine-tuning on organization- and developer-specific datasets. Our results show that there is a boost in prediction capabilities provided by both an organization-specific and a developer-specific additional fine-tuning, with the former being particularly performant. Such a finding generalizes across (i) the two subject organizations (i.e., Apache and Spring) and (ii) models of completely different magnitude (from 60M to 7B trainable parameters). Finally, we show that DL models fine-tuned on an organization-specific dataset achieve the same completion performance of pre-trained code models used out of the box and being sim10times larger, with consequent savings in terms of deployment and inference cost (e.g., smaller GPUs needed). |
|
2025-03-21T00:00:00 | 2503.13891 | Where do Large Vision-Language Models Look at when Answering Questions? | [
"Xiaoying Xing",
"Chia-Wen Kuo",
"Li Fuxin",
"Yulei Niu",
"Fan Chen",
"Ming Li",
"Ying Wu",
"Longyin Wen",
"Sijie Zhu"
] | https://github.com/bytedance/LVLM_Interpretation | Large Vision-Language Models (LVLMs) have shown promising performance in vision-language understanding and reasoning tasks. However, their visual understanding behaviors remain underexplored. A fundamental question arises: to what extent do LVLMs rely on visual input, and which image regions contribute to their responses? It is non-trivial to interpret the free-form generation of LVLMs due to their complicated visual architecture (e.g., multiple encoders and multi-resolution) and variable-length outputs. In this paper, we extend existing heatmap visualization methods (e.g., iGOS++) to support LVLMs for open-ended visual question answering. We propose a method to select visually relevant tokens that reflect the relevance between generated answers and input image. Furthermore, we conduct a comprehensive analysis of state-of-the-art LVLMs on benchmarks designed to require visual information to answer. Our findings offer several insights into LVLM behavior, including the relationship between focus region and answer correctness, differences in visual attention across architectures, and the impact of LLM scale on visual understanding. The code and data are available at https://github.com/bytedance/LVLM_Interpretation. |
2025-03-21T00:00:00 | 2503.07906 | Painting with Words: Elevating Detailed Image Captioning with Benchmark and Alignment Learning | [
"Qinghao Ye",
"Xianhan Zeng",
"Fu Li",
"Chunyuan Li",
"Haoqi Fan"
] | Image captioning has long been a pivotal task in visual understanding, with recent advancements in vision-language models (VLMs) significantly enhancing the ability to generate detailed image captions. However, the evaluation of detailed image captioning remains underexplored due to outdated evaluation metrics and coarse annotations. In this paper, we introduce DeCapBench along with a novel metric, DCScore, specifically designed for detailed captioning tasks. DCScore evaluates hallucinations and fine-grained comprehensiveness by deconstructing responses into the smallest self-sufficient units, termed primitive information units, and assessing them individually. Our evaluation shows that DCScore aligns more closely with human judgment than other rule-based or model-based metrics. Concurrently, DeCapBench exhibits a high correlation with VLM arena results on descriptive tasks, surpassing existing benchmarks for vision-language models. Additionally, we present an automatic fine-grained feedback collection method, FeedQuill, for preference optimization based on our advanced metric, showing robust generalization capabilities across auto-generated preference data. Extensive experiments on multiple VLMs demonstrate that our method not only significantly reduces hallucinations but also enhances performance across various benchmarks, achieving superior detail captioning performance while surpassing GPT-4o. |
Subsets and Splits