Markov Models
Counterfactual Conservative Q Learning for Offline Multi-agent Reinforcement Learning Jianzhun Shao, Y un Qu
MARL in real scenarios is still challenging due to the same safety and efficiency concerns in single-agent setting, then it is worth conducting investigation for offline RL in multi-agent setting.
Spectral Entry-wise Matrix Estimation for Low-Rank Reinforcement Learning
Both algorithms yield state-of-the-art performance guarantees.
Multi-Reward Best Policy Identification
This bound guides the design of an optimal exploration policy attaining minimal sample complexity. However, this lower bound involves solving a hard non-convex optimization problem.
Reinforcement learning from Human Feedback (RLHF) learns from preference signals, while standard Reinforcement Learning (RL) directly learns from reward
The latter case can be further reduced to adversarial MDP when preferences only depend on the final state.
DIFFER: Decomposing Individual Reward for Fair Experience Replay in Multi-Agent Reinforcement Learning
The code is available in https://github.com/cathyhxh/DIFFER.
OPERA: Automatic Offline Policy Evaluation with Re-weighted Aggregates of Multiple Estimators
Unfortunately, choosing the best OPE algorithm for each task and domain is still unclear.
Integrating Suboptimal Human Knowledge with Hierarchical Reinforcement Learning for Large-Scale Multiagent Systems
While agents' learning is data-driven, sampling from millions