Arpan Mukherjee

I am an Informed-AI postdoctoral researcher with Prof. Deniz Gündüz at the IPC lab in Imperial College London, where I work on the theory of language models. I obtained my Ph.D. degree at the Department of Electrical, Computer and Systems Engineering (ECSE), Rensselaer Polytechnic Institute (RPI), advised by Prof. Ali Tajer. Prior to joining RPI, I spent two wonderful years at the Indian Institute of Technology, Kharagpur, where I was advised by Prof. Mrityunjoy Chakraborty.

I am broadly interested in problems at the intersection of signal processing, statistics, and machine learning. Much of my research lies in the paradigm of sequential experiental design, where I have looked into various algorithmic aspects of multi-armed bandits. The overarching theme of my research is to investigate inference problems in an adaptive and sample-efficient setting. Phrases that attract my attention include optimal stopping, active learning, data-efficient decision making and identification problems in the context of experimental design. I am also keenly interested in algorithmic facets such as robustness and risk-sensitivity. Recently, I have started looking into reinforcement learning from human feedback (RLHF), where I am excited about sample complexity, preference diversity, and multi-objective RLHF.