Reinforcement learning · MLD-0833
Off-policy evaluation
Estimating a target policy's performance using data collected by another policy.
Original reference page · detailed expansion pending. Original definitions, examples, notes and citations are preserved below. New examples are illustrative, not benchmark results.
Original example
Evaluate a new ranking policy from logged interactions.
Original caveat
Poor action overlap can make estimates unreliable.
Context and aliases
Level: Advanced. Stages: Theory · Training · Evaluation.
Tags: RL · DL · SLM · LLM.
Aliases: OPE.
Related concepts
Concepts linking here
Original references
- OpenAI Spinning Up: reinforcement learning ↗ · Consulted
- TRL documentation ↗ · Consulted
- Original reference URL ↗