When your dataset lies to your face. (Data Poisoning Python)
AI & ML interests
None defined yet.
Recent Activity
Organization Card
✨ TQRG - The Quasar Research Group
Welcome to TQRG - The Quasar Research Group, led by Rui Abreu. We are a passionate, interdisciplinary team tackling challenges in software testing, debugging, security, energy efficiency, and machine learning.
Our mission is to create cutting-edge tools and methodologies that enhance software quality, performance, and sustainability. 🚀
👥 Members
- Rui Abreu — Group Leader
- Alexandre Perez
- Thomas Durieux
- Luis Cruz
- Sofia Reis
- Bruno Sotto-Mayor
- João Ferreira
- Ricardo Martins
- Francisco Ribeiro
- Nuno Sabino
- David Coimbra
- Eduard Pinconschi
- Cláudia Mamede
- Luís Gomes
- Rui Melo
🧪 Projects
Project | Description |
---|---|
GZoltar | Automatic Testing & Debugging using Spectrum-based Fault Localization (SFL) |
GZoltar-plugin | VS Code plugin with all GZoltar commands integrated |
Physalia | Energy measurement framework for mobile apps |
Leafactor | Automatic refactoring tool for improving energy efficiency in Android apps |
Energy Patterns | Open database of energy-related design patterns for mobile apps (iOS & Android) |
Android Test Inspector | Android App Hall of Fame |
Pangolin | Software visualization toolset for feature localization |
SecBench | Database of real-world security vulnerabilities |
SECOM | Convention for Security Commit Messages |
VDET | Transformer-based VS Code extension for detecting Java vulnerabilities |
Tandem-FL | Test minimizer based on Mozilla’s Lithium tool |
📚 Selected Publications
📘 2023
- Gabriela K Michelon et al. Spectrum-based feature localization for families of systems. JSS.
- Francisco Ribeiro et al. GPT-3-Powered Type Error Debugging. SLE'23.
📘 2022
- Cláudia Mamede et al. Transformer-based IDE plugin for vulnerability detection. ASE'22.
- Sofia Reis et al. Improving a Security Linter with Feedback. ASE'22.
- 🏆 MSR 2022 Foss Impact Paper Award: SECOM: Security Commit Message Convention. MSR'22.
- Eduard Pinconschi et al. Maestro: Program Repair Benchmarking Platform. ISSTA'22.
- Francisco Ribeiro et al. Framing Program Repair as Code Completion. APR'22.
📘 2021
- David Coimbra et al. Detecting C Security Vulnerabilities with Source Code Representations. DX'21.
- Sofia Reis et al. Fixing vulnerabilities potentially hinders maintainability. EMSE.
- Rui Abreu et al. Reducing Time-To-Fix for Fuzzer Bugs. ASE'21.
📘 2020–2018
- Bruno Sotto-Mayor et al. Smell-Based Defect Prediction with Designite. DX'20.
- Thomas Durieux et al. Empirical Study of Flaky Builds on Travis CI. MSR'20.
- Luis Cruz & Rui Abreu. Energy Pattern Catalog. EMSE.
- Alexandre Perez et al. Qualitative Reasoning for Fault Localization. IJCAI / ICSE.
📘 2017 and earlier
- Sofia Reis & Rui Abreu. SECBENCH: A Database of Security Vulnerabilities. ESORICS'17.
- Luis Cruz et al. Leafactor: Automatic Refactoring for Energy. MOBILESoft'17.
- Alexandre Perez et al. Test-Suite Diagnosability Metric. ICSE'17.
For the full list of publications, visit our Google Scholar profiles or institutional page.
🤝 Join Us / Get in Touch
We welcome collaborations, student researchers, and industry partners. Whether you're curious about energy-efficient apps, AI-powered debugging, or software security — we’d love to hear from you!
📫 Email: rui [at] computer.org
“Shining light on the dark matter of software systems.” ✨
-
A ground-truth dataset of real security patches
Paper • 2110.09635 • Published -
Are Sparse Autoencoders Useful for Java Function Bug Detection?
Paper • 2505.10375 • Published -
TQRG/devign_gpt2_tokenized
Viewer • Updated • 47.9k • 432 -
TQRG/bigcode_stack_dedup_python_gpt2_tokenized
Viewer • Updated • 5.05M • 134
When your dataset lies to your face. (Data Poisoning Python)
-
A ground-truth dataset of real security patches
Paper • 2110.09635 • Published -
Are Sparse Autoencoders Useful for Java Function Bug Detection?
Paper • 2505.10375 • Published -
TQRG/devign_gpt2_tokenized
Viewer • Updated • 47.9k • 432 -
TQRG/bigcode_stack_dedup_python_gpt2_tokenized
Viewer • Updated • 5.05M • 134
models
0
None public yet
datasets
11
TQRG/trust_issues_100k_6percent
Viewer
•
Updated
•
100k
•
44
TQRG/trust_issues_100k_3percent
Viewer
•
Updated
•
100k
•
44
TQRG/trust_issues_10k_3percent
Viewer
•
Updated
•
10k
•
49
TQRG/trust_issues_10k_6percent
Viewer
•
Updated
•
10k
•
45
TQRG/trust_issues_1k_6percent
Viewer
•
Updated
•
1k
•
47
TQRG/trust_issues_1k_3percent
Viewer
•
Updated
•
1k
•
45
TQRG/reset23_stack_java_gpt2_tokenized
Viewer
•
Updated
•
3.53M
•
103
TQRG/bigcode_stack_dedup_python_gpt2_tokenized
Viewer
•
Updated
•
5.05M
•
134
TQRG/devign_gpt2_tokenized
Viewer
•
Updated
•
47.9k
•
432
TQRG/raw-osv-data
Viewer
•
Updated
•
24.2k
•
5