Student Hiring
I am currently hiring PhD students for Fall 2027 in the areas of Generative AI and AI Agent.
Strong skill sets in mathematical thinking are desired for the position.
Research Interests
ML algorithms
- Generative AI: vision-language models, diffusion models, synthetic data
- AI Safety: jailbreaking, AI alignment, RSI safety
- Robust ML: distribution shift, test-time adaptation, transferability
ML Applications
- Health sciences: clinical world model, RL for treatment optimization, frontier model for EHR
- Materials design: generative models, Bayesian optimization
Publications
You can find the list of publications from Google scholar.
News
Congrats to Xin Hu for passing the prospectus! (Sep 2026)
Our paper titled "UtoMe: Observation-Uncertainty-Guided Token Merging for Weather Foundation Models" has been accepted to NeurIPS (Sep 2026).
Our paper titled "Adapting in the Dark: Efficient and Stable Test-Time Adaptation for Black-Box Models" has been accepted to NeurIPS (Sep 2026).
I'm co-organizing a Workshop titled "Data4VFM: Data-Centric Learning for Visual and Multimodal Foundation Models" at WACV 2027 (Aug 2026).
We have been awarded the NIH grant titled "The SKIN DEEP Study: Advancing AI Diagnosis of Skin Diseases through Multi-Sensory Data Enhancement and Processing" (Aug 2026).
I'm serving as SAC for ICLR'27.
Our paper titled "Intervention-Aware Clinical World Model for Post–Op Outcome Forecasting in Cardiology" has been accepted to MICCAI Workhop MWM2026 (July 2026).
Our paper titled "Continual Test-Time Adaptation in Computer Vision: Methods, Benchmarks, and Future Directions" has been accepted to TMLR 2026.
Our paper titled "Less Tokens, Better Forecasts: Sparse Residual Routing for Efficient Weather Prediction" has been accepted to ECCV 2026.
Our paper titled "Auto-Prompting: Layer-Specific Prompt Fusion Discovery via Differentiable Search" has been accepted to ECCV 2026.
Our paper titled "Integrating transfer learning and uncertainty quantification for data-efficient materials discovery" was accepted for publication at Journal of Physics: Materials.
Welcome Rayyan to the Maching Learning Group!
Our paper titled "Stop Comparing LLM Agents Without Disclosing the Harness" was accepted to ICML 2026 Workshop AIWILD.
Our paper titled "Novel Radiomics-Integrated Panel for Preoperative Stratification of Pancreatic Neuroendocrine Tumors (PNETs)" has been accepted for publication at Cancers-mdpi
Congrations Pavan Sanjay for receiving the MS degree in CS!
Our paper titled "Less Tokens, Better Forecasts: Sparse Residual Routing for Efficient Weather Prediction" was accepted to CVPR 2026 Workshop AI4RWC
Our paper titled "Mind the Gap: How Medical VLMs Underutilize Their Vision Encoders" was accepted to CVPR 2026 Workshop AI4RWC
Our paper titled "Visual Exclusivity Attacks: Automatic Multimodal Red Teaming via Agentic Planning" received the Best Short Paper Award at AIWILD @ ICLR2026!
Our paper titled "Agents in the Wild: Safety, Security, and the Illusion of Sociality on Moltbook" has been accepted for presentation at ICLR 2026 Workshop AIWILD (Mar 2026) arXiv:2602.13284
Our paper titled "Seeing Clearly, Reasoning Confidently: Plug-and-Play Remedies for Vision Langugae Model Blindness" has been accepted for publication at CVPR (Feb 2026).
Our paper titled "Alternative Reprogramming for Service Models" has been accepted for publication at CVPR (Feb 2026).
Congrats to Yunbei for successfully passing the prospectus! (Feb 2026).
Our paper titled "Prompt-based Adaptation in Large-scale Vision Models: A Survey" has been accepted for publication at TMLR (Jan 2026).
