Average-Reward Policy Identification from a Single Continuing Trajectory
For average-reward reinforcement learning, we develop an algorithm that learns where to collect information while ensuring that the required samples can be obtained efficiently.
ProofPaper ↗
Key points
- In reinforcement learning, an agent must explore an unknown environment to learn how to make good decisions.
- Unlike methods that can freely sample any state--action pair, an agent learning from a single trajectory must spend additional time navigating between states.
- We establish an upper bound on the number of interactions needed to identify a near-optimal policy.
- Furthermore, we prove a matching information theoretic lower bound, showing that, for a broad class of environments and at sufficiently high accuracy, learning from a single trajectory is no harder in the worst case than learning with unrestricted access to state action samples.
Sources (1)
- [1]Average-Reward Policy Identification from a Single Continuing TrajectoryarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 14, 04:57 PM
For average-reward reinforcement learning, we develop an algorithm that learns where to collect information while ensuring that the required samples can be obtained efficiently.
In reinforcement learning, an agent must explore an unknown environment to learn how to make good decisions.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 14, 2026Player-1: Learning to Play from Synthetic Experience
- Oct 14, 2026The Design Space of Tri-Modal Masked Diffusion Models
- Oct 14, 2026More Methods, Better Agents? A Reality Check on Self-Distillation in Agentic Reinforcement Learning
- Oct 14, 2026MAI-Gym: An Open All-in-One Platform for GUI Agent Evaluation and RL Training across Mobile and Computer Environments
- Oct 13, 2026Run fast, private agents on-device with LFM2.5-24B-A2B
- Oct 13, 2026LensErase: A Foundation Model for Erase Anything
