\Mathcal{o}}$ Revenue Regret

Revenue regret quantifies the difference between a decision-maker's actual revenue and the optimal revenue achievable in hindsight, often within dynamic environments like auctions or multi-agent reinforcement learning. Current research focuses on developing algorithms that minimize this regret under various constraints, such as limited policy updates (adaptivity constraints) or unknown market noise distributions, employing techniques like Thompson Sampling, Upper Confidence Bound (UCB) variations, and policy elimination. These advancements are significant for improving decision-making in online advertising, auction design, and other applications where sequential interactions and uncertainty are prevalent, leading to more efficient resource allocation and increased profitability.

Papers

February 5, 2024

Non-Stationary Latent Auto-Regressive Bandits
Anna L. Trella, Walter Dempsey, Finale Doshi-Velez, Susan A. Murphy
Multi Armed Bandit Non Stationary Environment Stochastic Reward Latent Bandit \Mathcal{o}}$ Revenue Regret

February 2, 2024

Near-Optimal Reinforcement Learning with Self-Play under Adaptivity Constraints
Dan Qiao, Yu-Xiang Wang
Multi Agent Reinforcement Learning Markov Game Self Play Optimal Reinforcement Learning Optimal Batch Adaptivity Constraint \Mathcal{o}}$ Revenue Regret

October 19, 2022

A Reinforcement Learning Approach in Multi-Phase Second-Price Auction Design
Rui Ai, Boxiang Lyu, Zhaoran Wang, Zhuoran Yang, Michael I. Jordan
Price Auction \Mathcal{o}}$ Revenue Regret Reserve Price

May 27, 2022

Lifting the Information Ratio: An Information-Theoretic Analysis of Thompson Sampling for Contextual Bandits
Gergely Neu, Julia Olkhovskaya, Matteo Papini, Ludovic Schwartz
Contextual Bandit Thompson Sampling Information Gain Bayesian Regret Logistic Bandit \Mathcal{o}}$ Revenue Regret

March 14, 2022

Learning Markov Games with Adversarial Opponents: Efficient Algorithms and Fundamental Limits
Qinghua Liu, Yuanhao Wang, Chi Jin
Markov Game Efficient Algorithm Zero Sum Game Adversarial Agent Fundamental Limit Game State \Mathcal{o}}$ Revenue Regret

\Mathcal{o}}$ Revenue Regret

Papers

Non-Stationary Latent Auto-Regressive Bandits

Near-Optimal Reinforcement Learning with Self-Play under Adaptivity Constraints

A Reinforcement Learning Approach in Multi-Phase Second-Price Auction Design

Lifting the Information Ratio: An Information-Theoretic Analysis of Thompson Sampling for Contextual Bandits

Learning Markov Games with Adversarial Opponents: Efficient Algorithms and Fundamental Limits