Reading Group @ SICE, IUB

Welcome to the homepage of our reading group for reinforcement learning.

Info

Host: Chao Tao

Time: bi-weekly on 1:00pm - 3:00pm

Place: Luddy Hall 3069

Schedule

• 1/17, Chao Tao on Introduction to Reinforcement Learning (notes)

Keywords

Markov Decision Process (MDP); Value Functions; Bellman Equations; State Occupancy; Q Learning;

• 1/31, Chao Tao on Minimax regret bounds for reinforcement learning (notes)

Keywords

Tabular Episodic MDP; Model-based; Optimistic Algorithm; Frequentist Regret; $\widetilde{\mathcal{O}}(H\sqrt{SAT})$;

• 2/14, Chao Tao on Is Q-learning Provably Efficient? (notes)

Keywords

Tabular Episodic MDP; Model-free; Optimistic Algorithm; Frequentist Regret; $\widetilde{\mathcal{O}}(H^2\sqrt{SAT})$;

• 2/28, Chao Tao on (More) Efficient Reinforcement Learning via Posterior Sampling (notes)

Keywords

Tabular Episodic MDP; Thompson Sampling; Bayesian Regret; $\widetilde{\mathcal{O}}(HS\sqrt{AT})$;

• 4/10, Chao Tao on Learning Unknown Markov Decision Processes: A Thompson Sampling Approach (notes)

Keywords

Tabular Infinite Undiscounted MDP; Weakly Communicating; Thompson Sampling; Bayesian Regret; $\widetilde{\mathcal{O}}(D'S\sqrt{AT})$;

• 5/1, Nikolai Karpov on Concurrent PAC RL