Search papers, labs, and topics across Lattice.
MR-Search is introduced as an in-context meta-RL formulation for agentic search, training a policy to adapt its search strategy across episodes by conditioning on past experiences. The approach incorporates explicit self-reflections generated after each episode as context to guide subsequent attempts, enhancing exploration. A multi-turn RL algorithm estimates dense relative advantages at the turn level for fine-grained credit assignment, leading to 9.2%-19.3% improvements over RL baselines across benchmarks.
Agentic search gets a meta-RL boost: MR-Search learns to self-reflect and adapt search strategies across episodes, significantly outperforming standard RL baselines.
This paper introduces MR-Search, an in-context meta reinforcement learning (RL) formulation for agentic search with self-reflection. Instead of optimizing a policy within a single independent episode with sparse rewards, MR-Search trains a policy that conditions on past episodes and adapts its search strategy across episodes. MR-Search learns to learn a search strategy with self-reflection, allowing search agents to improve in-context exploration at test-time. Specifically, MR-Search performs cross-episode exploration by generating explicit self-reflections after each episode and leveraging them as additional context to guide subsequent attempts, thereby promoting more effective exploration during test-time. We further introduce a multi-turn RL algorithm that estimates a dense relative advantage at the turn level, enabling fine-grained credit assignment on each episode. Empirical results across various benchmarks demonstrate the advantages of MR-Search over baselines based RL, showing strong generalization and relative improvements of 9.2% to 19.3% across eight benchmarks. Our code and data are available at https://github.com/tengxiao1/MR-Search.
AI2
UW