cross-entropy method
PulseAugur coverage of cross-entropy method — every cluster mentioning cross-entropy method across labs, papers, and developer communities, ranked by signal.
-
AI hybrid design enables efficient visual navigation for swarm robots
Researchers have developed a novel AI-driven method for designing decentralized controllers for swarm robotics, enabling autonomous visual navigation in complex environments. This approach combines multi-agent reinforce…
-
New 'Follow the Winners' algorithm enhances LLM reinforcement learning
Researchers have introduced "Follow the Winners" (FTW), a novel critic-free reinforcement fine-tuning algorithm designed for agentic large language models. Unlike existing GRPO-style methods that rely on impractical rep…
-
New research explores dual role of world models in cross-entropy method
A new research paper explores the dual role of world models within the cross-entropy method (CEM), highlighting their function in both selecting actions and refining future proposals. The study found that scoring errors…
-
Study explores conditional diffusion models for friction-prone control systems
Researchers have explored the use of conditional diffusion models for open-loop control, specifically focusing on systems with dry friction and stiction. Their study, "A Study of Conditional Diffusion Models for Open-Lo…
-
New methods enhance AI planning with latent world models · 4 sources tracked
Researchers have developed new methods to improve planning in latent world models, which are systems that predict outcomes of action sequences. One approach, Reinforced Planning (RP1), learns to improve multi-step plans…
-
New conservation law quantifies program discovery costs
A new paper introduces a conservation law for program discovery, suggesting that injecting structural knowledge into a search algorithm trades off directly against the search effort. This law quantifies the cost of find…
-
Robotics research uses neural beliefs for robust grasping under uncertainty
Researchers have developed a new method for robust dexterous grasping in robotics by employing variational neural belief parameterizations. This approach models uncertainty in contact parameters and object pose using a …