I am a fourth-year Ph.D. student at the Manning College of Information & Computer Sciences at UMass Amherst, advised by Amir Houmansadr. My general research interest lies in trustworthy machine learning, with a current focus on various aspects of trustworthy generative AI, including large language models and multimodal models. I obtained my bachelor’s degree from Sharif University of Technology.
News
- February 5, 2025: Our new preprint, “OverThink: Slowdown Attacks on Reasoning LLMs”, is out.
- February 1, 2025: Our new preprint, “Riddle Me This! Stealthy Membership Inference for Retrieval-Augmented Generation”, is out.
- January 20, 2025: Our new preprint, “Synthetic Data Can Mislead Evaluations: Membership Inference as Machine Text Detection”, is out.
- September 16, 2024: 🔦 Our paper, “Iteratively Prompting Multimodal LLMs to Reproduce Natural and AI-Generated Images”, was selected for an oral spotlight presentation at the 2024 COLM (top 2%).
- August 30, 2024: Our paper, “Diffence: Fencing Membership Privacy With Diffusion Models”, was accepted at the 2025 NDSS.
- July 10, 2024: Our paper, “Iteratively Prompting Multimodal LLMs to Reproduce Natural and AI-Generated Images”, was accepted at the 2024 COLM.
- May 20, 2024: I started a new position as a Research Intern at Oracle, where I am working on membership inference attacks in LLMs.
- May 17, 2024: I received my Master’s degree from, and subsequently advanced to PhD candidate.
- May 6, 2024: 🌟 I received the Distinction Award for Portfolio Excellence from the Manning College of Information & Computer Sciences at the University of Massachusetts Amherst.
- March 15, 2024: I presented a talk titled “Stealing the decoding algorithms of language models”, at the 2024 New England Security Day.
- December 22, 2023: Our paper, “Memory Triggers: Unveiling Memorization in Text-To-Image Generative Models through Word-Level Duplication”, was accepted at the 2024 PPAI.
- November 30, 2023: 🏆 Our paper, “Stealing the decoding algorithms of language models”, won a Distinguished Paper Award at the 2023 CCS.
- May 8, 2023: Our paper, “Stealing the decoding algorithms of language models”, was accepted at the 2023 CCS.