Upcoming Events

Machine Learning Seminar Spring 2026 | Simplifying AI Models with optimization and statistics

Abstract: Highly performant AI models such as Large Language Models (LLMs) have achieved remarkable performance across various domains. It is widely acknowledged that their large model sizes lead to high computational costs (storage, inference latency, memory, etc) making serving and deployment expensive especially in low-resource environments. LLM pruning or more generally model compression is a line of work where LLM model parameters are compressed to reduce model footprint. For example, post-training model pruning aims to achieve model compression by removing less-important parameters while retaining model utility as much as possible. Depending upon available hardware, different types of model compression methods can be useful (eg, sparsity, sparse plus low-rank, quantization, etc). These problems can be formulated as large-scale discrete optimization problems posing interesting algorithmic research questions. In this talk, I will discuss our recent experience in using tools and insights from high-dimensional optimization and statistics in the context of LLM compression.

Bio: Rahul Mazumder is the NTU Associate Professor of Operations Research and Statistics at MIT Sloan School of Management. He is affiliated with MIT OR Center, MIT Center for Statistics and Data Science, LIDS and IDSS. His research interests are at the intersection of statistics, machine learning and mathematical programming, and their applications to industry, the government, and the sciences. He is a recipient of the Leo Breiman Junior Award from the American Statistical Association, International Indian Statistical Association Early Career Award in Statistics and Data Science, INFORMS Donald P. Gaver, Jr. Early Career Award for Excellence in Operations Research, INFORMS Optimization Society Young Researchers Prize, Office of Naval Research Young Investigator Award, INFORMS ICS Prize (Honorable Mention). He is currently serving as an AE of the Annals of Statistics, Operations Research and Journal of Machine Learning Research, and was an AE at Bernoulli.