Shubhanshu Shekhar, MICS Ph.D candidate from the department of Electrical and Computer Engineering has been awarded the prestigious Shannon Graduate Fellowship for the academic year 2019-2020. This annual award was created by Prof. Jack K. Wolf in 2008 and has historically recognized a distinguished group of graduate student at UCSD who, at UCSD and beyond, have demonstrated outstanding achievements in information theory. As stated in the fellowship announcement, the award recognizes Shekhar’s “exceptional quality of his technical accomplishments” with his PhD advisor and MICS director Prof Tara Javidi as well his senior collaborator and MICS faculty Prof. Siavash Mirarab.
“My research interests lie in the general area of sequential decision making under uncertainty, where an agent interacts sequentially with an unknown environment in order to achieve a specified objective. As the environment is unknown, the agent is faced with the dilemma between exploitation and exploration, i.e., acting optimally based on current knowledge of the environment or selecting actions which are most informative about the environment. The goal here is to design strategies for the agent which have low regret, i.e., sub-optimality w.r.t. a hypothetical agent who acts optimally with the full environment information.
The general framework described above models several important problems such as sequential optimization, active learning, reinforcement learning and active hypothesis testing; and forms the theoretical and algorithmic foundation for diverse applications such as AutoML, product recommendation, opinon-polling and automated diagnostic systems. In my Ph.D. research, I have worked on designing adaptive information acquisition strategies for sequential learning and optimization problems, and rigorously demonstrating the resulting theoretical and/or empirical benefits over competing baselines.”
Congratulations to Shubhanshu on his award and we wish him the best of luck on his research!
