Last-Iterate Convergence with Full and Noisy Feedback in Two-Player Zero-Sum Games.
Kenshi AbeKaito AriuMitsuki SakamotoKentaro ToyoshimaAtsushi IwasakiPublished in: AISTATS (2023)
Keyphrases
- perfect information
- optimal strategy
- imperfect information
- convergence rate
- user feedback
- machine learning
- convergence speed
- game theoretic
- reinforcement learning algorithms
- relevance feedback
- decision problems
- real time
- feedback loop
- data sets
- dynamic programming
- global convergence
- nash equilibria
- initial conditions
- single agent
- multi agent
- incomplete data
- image retrieval
- user interaction