Niharika (Ari) Sravan
Assistant Professor of Physics, Drexel University
I am interested in the use of machine learning techniques to streamline scientific discovery and inference by addressing challenges of scale, speed or complexity. I am intrigued by the opportunities presented by the Rubin observatory to study transients spanning novel but challenging phase spaces. Scientifically, I am especially interested in sources that are multimessenger counterparts. Recently, I have focused on the design of systems that enable real-time decision-making in follow-up, for identifying and characterizing transients discovered by surveys. Strategizing an optimal follow-up program requires adaptive sequential decision-making given evolving light curve data that maximizes a global objective despite incomplete information and is robust to stochasticity introduced by detectors/observing conditions. I use reinforcement learning that allows artificial agents to implicitly learn the physics/detector dynamics and the behavior policy that maximize a designated objective through experience. Such systems go beyond classifiers by soliciting and adapting to new information, closing the current gap between data and science decisions. It also goes beyond active learning techniques by computing the explore-exploit tradeoff: is it better to act on the current best guess v. trying something less optimal to improve this estimate and later act on the greedy estimate.
As a mentor I prioritize open and honest communication with my mentees. Recognizing the importance of networking and collaborations for career outcomes, I ensure that students are aware of recent developments in research and have the opportunity to meet and interact with the relevant members of the community. I make sure to highlight opportunities for research collaboration and any long term strategic benefits of investing in them. When mentoring my students from underrepresented backgrounds, I am particularly sensitive to any internalized oppression and deconstructing it. I am open about personal issues I faced through my career as an immigrant woman of color in STEM. My goal is for them to understand these issues and find ways to manage their experiences to get the outcomes they want from their lives. A key goal is to ensure the mentee is aware of career and funding opportunities and supporting them in writing their applications for them. Personally, I keep up-to-date with developments in core AI research and progress in industry applications and ensure that my students are aware of these.