Meta-Reinforcement Learning with Self-Reflection for Agentic Search

TL;DR

This paper presents MR-Search, a meta-reinforcement learning approach that uses self-reflection to improve agent exploration across episodes, leading to better performance on multiple benchmarks.

Contribution

Introduces MR-Search, a novel meta-RL framework that incorporates self-reflection and cross-episode learning for enhanced agent exploration and adaptation.

Findings

Achieves 9.2% to 19.3% improvements over baselines

Demonstrates strong generalization across benchmarks

Utilizes a multi-turn RL algorithm for fine-grained credit assignment

Abstract

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.