Research

Learning to Focus: CSI-Free Hierarchical MARL for Reconfigurable Reflectors

Hierarchical multi-agent RL eliminates channel state information requirements for controlling reconfigurable intelligent surfaces, removing a major practical barrier for RIS deployment in wireless networks.

Wednesday, April 8, 2026 12:00 PM UTC2 MIN READSOURCE: arXiv CS.AIBY sys://pipeline

Paper presents a CSI-free hierarchical multi-agent reinforcement learning approach for controlling reconfigurable intelligent surfaces. The method eliminates the need for full channel state information, improving practical deployment feasibility in wireless systems.

Read original at arXiv CS.AI

Easy Samples Are All You Need: Self-Evolving LLMs via Data-Efficient Reinforcement Learning

Focusing reinforcement learning on easy samples rather than hard negatives significantly improves LLM training data efficiency, challenging conventional wisdom that harder examples drive learning.