|
Invited talks
Upcoming talks
Statistical theory for reinforcement learning: Oracle inequalities, Markov chains, and stochastic approximation, Young Researchers Workshop, Cornell University, Oct 2022
Optimal and Instance-dependent Guarantees for Markovian Linear Stochastic Approximation, APS student paper competition at INFORMS Annual Meeting, Indianapolis, Oct 2022
On The Statistical Complexity Of Reinforcement Learning With Function Approximation, INFORMS Annual Meeting, Indianapolis, Oct 2022
Rethinking semi-parametric efficiency for off-policy estimation: a non-asymptotic perspective, BLISS seminar, UC Berkeley, Oct 2022
Optimal variance-reduced stochastic approximation in Banach spaces, Applied and Computational Math seminar, Georgia Institute of Technology, Nov 2022
Past talks
Optimal algorithms for reinforcement learning: Oracle inequalities, Markov chains, and stochastic approximation, International Conference on Continuous Optimization, Lehigh University, July 2022
Statistical theory for reinforcement learning: Oracle inequalities, Markov chains, and stochastic approximation, Neyman seminar, Department of Statistics, UC Berkeley, January 2022
High-Order Langevin diffusion yields an accelerated MCMC algorithm, Simons Institute program on Geometric Methods in Optimization and Sampling, October 2021
Optimal oracle inequalities for projected fixed-point equations, Machine Learning reading group, UT Austin, September 2021
On the sample complexity of reinforcement learning with policy space generalization, Simons Institute, reinforcement learning theory student seminar, December 2020
Langevin diffusions in modern statistical learning: discretization, sampling, and posterior concentration, statistics reading group, University of Wisconsin, Madison, July 2020
High-Order Langevin diffusion yields an accelerated MCMC algorithm, ACO student seminar, Georgia Institute of Technology, October 2019
Generalization bounds of SGLD for non-convex learning: two theoretical viewpoints, Simons Institute, deep learning student seminar, October 201
|