Chinmaya Kausik

• Mathematical problems motivated by tangible, real-world questions. These days, my work focuses on sequential decision making with offline data. I also have side projects in deep learning and control theory, and I intend to combine my current interests with ideas in differential privacy very soon. On the other hand, a lot of my undergraduate background was in geometry, topology and dynamics, with work in computer-assisted topology and geometry.
• Increasing accessibility to and in higher mathematics and creating communities where ideas cross pollinate and people pull each other up. This is part of why I co-founded and co-organize Monsoon Math Camp - an online math camp for promising high school students introducing them to advanced undergraduate and graduate-level math. You should check it out! I have also been involved in building and expanding other mathematical communities, like platforms for the PolyMath REU, DRP programs and the undergraduate math organization at IISc, etc.

• Working on projects on offline RL in confounded tabular MDPs and learning mixtures of MDPs.
• Thinking about minimax optimal algorithms for offline policy evaluation (OPE) and the role of the geometry of action sets.
• Working on double descent in denoising under the guidance of Rishi Sonthalia. Work continued from MREG 2022.
• Continuing work on our project from LOGML 2022! I was a participant in Dr. Eli Meirom’s group, planning to work on using RL for graph rewiring in GNNs to prevent oversquashing for long range problems.
• Mentoring undergraduate students on a project about using machine learning to enhance model predictive control.
• Fleshing out ideas for more academic communities like Monsoon Math.

primary goals

• Work on learning other mixtures of time series with control input.
• Work on generating synthetic data privately for time series, starting with MDPs and Markov Chains.
• Learn about differential privacy and its intersection with sequential decision-making.
• Start maintaining my progress log again.
• Learn about safe RL and think about techniques beyond primal-dual ones, perhaps using model-based RL with uncertain models.
• Watch lectures from the Data Driven Decision Processes program at the Simons Institute this semester.
• Gain a comprehensive view of techniques that go into minimax lower bounds in RL.

side-quests

• The theory behind GNNs and deep learning in general.
• Causality and its interaction with sequential decision making and RL.
• Algorithmic fairness.
• Geometric and topological insights for data analysis and machine learning (for example, non-positive curvature representation learning).
• Natural Language Processing.
• Using insights from machine learning for biology. In a specific example, learning a hierarchical or causal structure from genomics