معرفی
Mohammad Mahdi Khalili is an Assistant Professor in the Department of Computer Science and Engineering at The Ohio State University's College of Engineering. He also serves as a part-time Research Scientist at Yahoo! Research, focusing on theoretical and applied machine learning with emphasis on robustness, interpretability, and model compression for large-scale systems including LLMs.
Dr. Khalili earned his Ph.D. in Electrical Engineering and Computer Science and MSc in Applied Mathematics from the University of Michigan, Ann Arbor. Prior to OSU, he was a research scientist at Yahoo Research and a postdoctoral researcher at UC Berkeley.
His research spans mechanistic interpretability of foundation models, privacy-aware model compression, and fairness in sequential decision-making. Current projects include counterfactual reasoning for fair ML, physiological signal analysis for worker health monitoring, and corruption analysis in model mechanisms. His work bridges theoretical guarantees with practical healthcare and security applications.
Recent publications reveal strong focus on model compression techniques (block-wise sparsity, quantization) and interpretability methods for LLMs, with growing emphasis on physiological signal processing applications. The 2024-2025 papers show increasing integration of healthcare domains like ECG analysis and construction worker monitoring.
Dr. Khalili currently advises four PhD students:
- Zhiqun Zuo: Counterfactual Reasoning
- Zhongteng Cai: Privacy-Aware Model Compression (UAI travel grant recipient)
- Ding Zhu: Trustworthy Model Compression & Time Series Analysis
- Vishnu Chhabra: Mechanistic Interpretability
He actively contributes to the ML community through invited talks (Midwest Machine Learning Symposium) and publications in top venues including NeurIPS, ICML, and UAI.
