Matt Fredrikson
Associate Professor
Studies the security, privacy and reliability of machine learning, including adversarial attacks on language models.
Matt Fredrikson's lab at CMU works on trustworthy & safe ML and large language models. 21 papers since 2023, 16 at top venues, mostly ICLR and NeurIPS. 9 alumni. Applications close .
At a glance
- papers, 2023–2026
- 5/ year
- at T1 venues
- 76%
- PhD students
- 0
- MS students
- 0
- alumni listed
- 9
Papers by yearfrom the lab homepage, 2023–2026
- 6 papers2026 · T1 5 · collecting
- 6 papers2025 · T1 5
- 6 papers2024 · T1 3
- 3 papers2023 · T1 3
T1 venuesOtherVenues: ICLR, NeurIPS, ACL, CHI, CVPR
Research focuspaper titles that use each topic's words
- Large language models10 papers
- attacks3 papers
- Trustworthy & safe ML1 paper
- adversarial1 paper
- reliability0 papers
Paper list
From the lab homepage.
20266 papers
- ICLR (T1)Comparing AI Agents to Cybersecurity Professionals in Real-World Penetration Testing (opens in a new tab)
- ACL (T1)Jailbreak-Zero: A Path to Pareto Optimal Red Teaming for Large Language Models (opens in a new tab)
- ICLR (T1)LipNeXt: Scaling up Lipschitz-based Certified Robustness to Billion-parameter Models (opens in a new tab)
- CVPR (T1)Omni-Attack: Adversarial Attacks on Open-Ended VQA in Black-Box Multimodal LLMs (opens in a new tab)
- NDSS (T1)PrivCode: When Code Generation Meets Differential Privacy (opens in a new tab)
- ICISSPSequential Pattern Recognition Attacks against Deployed Topic-Based Mechanisms (opens in a new tab)
20256 papers
- ICLR (T1)AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents (opens in a new tab)
- ICLR (T1)Aligned LLMs Are Not Aligned Browser Agents (opens in a new tab)
- CHI (T1)LLM Whisperer: An Inconspicuous Attack to Bias LLM Responses (opens in a new tab)
- TrustComPrototype-Integrated Representation Learning for Novelty Detection (opens in a new tab)
- NeurIPS (T1)Safety Pretraining: Toward the Next Generation of Safe AI (opens in a new tab)
- NeurIPS (T1)Security Challenges in AI Agent Deployment: Insights from a Large Scale Public Competition (opens in a new tab)
20246 papers
- ICLR (T1)A Recipe for Improved Certifiable Robustness (opens in a new tab)
- NeurIPS (T1)Efficient LLM Jailbreak via Adaptive Dense-to-sparse Constrained Optimization (opens in a new tab)
- NeurIPS (T1)Improving Alignment and Robustness with Circuit Breakers (opens in a new tab)
- CVPRFIs Your Text-to-Image Model Robust to Caption Noise? (opens in a new tab)
- International Conference on Human Factors in Computing SystemsSales Whisperer: A Human-Inconspicuous Attack on LLM Brand Recommendations (opens in a new tab)
- KBSETransfer Attacks and Defenses for Large Language Models on Coding Tasks (opens in a new tab)
Where alumni went9 listed
Apply to CMUFall 2027
CMU SCS Computer Science PhD
- Round
- Single round
- Opens
- GRE
- See program page
- English
- TOEFL / IELTS / Duolingo (F-1/J-1 non-native speakers)
- Fee
- USD 100 (USD 80 by early deadline)
Confirmed on the program page.
Similar labsnearest by research summary
- DETUM
Stefanie Jegelka
- Learning theory & optimization
- SGNTU
Yeshwanth Cherapanamjeri
- Learning theory & optimization
- USUCSD
Jun-Kun Wang
- Learning theory & optimization
- GBOxford
Paul Goldberg
- Learning theory & optimization
- CNZhejiang
Tianhang Zheng
- Trustworthy & safe ML
- USUC Berkeley
Dawn Song
- Trustworthy & safe ML
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
CMUcloses