technique

decision-content decoupled reinforcement learning

techniqueactivedecision-content-decoupled-reinforcement-learning-7409a7d1·1 events·first seen 27d ago

Aliases: decision-content decoupled reinforcement learning

Co-occurring entities

web navigation benchmark Mem-π large language model agents episodic memory retrieval

More like this (12)

decoupled reinforcement learning Constrained Reinforcement Learning Decision Transformer self-play reinforcement learning Goal-Conditioned Reinforcement Learning Hierarchical Reinforcement Learning shielded reinforcement learning Conformal Decision Theory Reinforcement Learning Q-learning Soft Q-Learning AI-driven constraint reasoning

Recent events (1)

6arXiv · cs.CL·27d ago·source ↗

Mem-π: Adaptive Memory for LLM Agents via On-Demand Generation and Decoupled RL

Mem-π introduces a framework where a dedicated language or vision-language model generates context-specific guidance for LLM agents on demand, rather than retrieving static entries from episodic memory banks. The system is trained with a decision-content decoupled reinforcement learning objective that jointly learns when to generate guidance and what to generate, enabling abstention when generation would not help. Evaluated across web navigation, terminal-based tool use, and text-based embodied interaction benchmarks, Mem-π achieves over 30% relative improvement on web navigation tasks compared to retrieval-based and prior RL-optimized memory baselines.

Evaluation and Benchmarking Agent and Tool Ecosystem web navigation benchmark Mem-π large language model agents +3 more