Meta-RL Induces Exploration in Language Agents Reinforcement learning (RL) has enabled the training of large language model (LLM) agents to interact with the environment and to solve multi-turn long-horizon tasks. However, the RL-trained agents…

LaMer brings meta-RL to LLM agents: cross-episode credit + in-context reflection = stronger exploration, better pass@3 & OOD generalization across Sokoban, Minesweeper, Webshop, ALFWorld. Paper: arxiv.org/abs/2512.16848 #MetaRL #LLMAgents #ReinforcementLearning

Directed-MAML boosts efficiency in meta‑reinforcement learning

Directed‑MAML was tested on CartPole‑v1, LunarLander‑v2 and a two‑vehicle intersection scenario, achieving better performance than standard MAML with faster training. Read more: getnews.me/directed-maml-boosts-eff... #directedmaml #metarl

Meta‑RL Solution for Capacity‑Aware Scheduling in Multi‑Agent MDPs

A meta‑RL framework for capacity‑constrained multi‑agent MDPs was evaluated on industrial robots with limited repair technicians; the preprint was posted 26 September 2025. getnews.me/meta-rl-solution-for-cap... #metarl #robotmaintenance