How Google’s 'internal RL' could unlock long-horizon AI agents

Google researchers introduce ‘Internal RL,’ a technique that steers an models' hidden activations to solve long-horizon tasks ...
AZoRobotics

Simulation-Only Training Enables Quadruped Robots to Master Real-World Terrain

A quadruped robot uses deep reinforcement learning to master walking on varied terrains, demonstrating energy-efficient and ...
IEEE

Deep Reinforcement Learning for Online Reconfiguration of Active Distribution Network

Abstract: The operation and control of active distribution networks (ADNs) are becoming increasingly important due to the high penetration of renewable energy (RE). The inherent uncertainty of RE can ...
EurekAlert!

Multi-constraint reinforcement learning in complex robot environments

FPMCO decomposes multi-constraint RL into KL-projection sub-problems, achieving higher reward with lower computing than second-order rivals on the new SCIG robotics benchmark.
Tech Xplore

Reinforcement learning accelerates model-free training of optical AI systems

Optical computing has emerged as a powerful approach for high-speed and energy-efficient information processing. Diffractive ...
Frontiers

Reinforcement learning control method for greenhouse vegetable irrigation driven by dynamic clipping and negative incentive mechanism

Greenhouse vegetable production was a complex agricultural system influenced by multiple interrelated environmental and management factors. Its irrigation control was a critical but not singularly ...
blockchain

DeepMind Unveils AI System That Discovers Novel Reinforcement Learning Algorithms, Surpassing Human Designs

According to God of Prompt on Twitter, DeepMind has published groundbreaking research in Nature led by David Silver, introducing an AI meta-learning system capable of autonomously discovering entirely ...
marktechpost

Weak-for-Strong (W4S): A Novel Reinforcement Learning Algorithm that Trains a weak Meta Agent to Design Agentic Workflows with Stronger LLMs

W4S operates in turns. The state contains task instructions, the current workflow program, and feedback from prior executions. An action has 2 components, an analysis of what to change, and new Python ...
TechCrunch

The reinforcement gap — or why some AI skills improve faster than others

AI coding tools are getting better fast. If you don’t work in code, it can be hard to notice how much things are changing, but GPT-5 and Gemini 2.5 have made a whole new set of developer tricks ...