GigaBrain-0.5M*: a VLA That Learns From World Model-Based Reinforcement Learning
Paper • 2602.12099 • Published • 61
Collector 2
-
-
When to Memorize and When to Stop: Gated Recurrent Memory for Long-Context Reasoning
Paper • 2602.10560 • Published • 30 -
G-LNS: Generative Large Neighborhood Search for LLM-Based Automatic Heuristic Design
Paper • 2602.08253 • Published • 26 -
ROCKET: Rapid Optimization via Calibration-guided Knapsack Enhanced Truncation for Efficient Model Compression
Paper • 2602.11008 • Published • 18 -
SkillRL: Evolving Agents via Recursive Skill-Augmented Reinforcement Learning
Paper • 2602.08234 • Published • 72 -
MetaClaw: Just Talk -- An Agent That Meta-Learns and Evolves in the Wild
Paper • 2603.17187 • Published • 131 -
Generation Models Know Space: Unleashing Implicit 3D Priors for Scene Understanding
Paper • 2603.19235 • Published • 92 -
Memento-Skills: Let Agents Design Agents
Paper • 2603.18743 • Published • 53 -
MHPO: Modulated Hazard-aware Policy Optimization for Stable Reinforcement Learning
Paper • 2603.16929 • Published • 12 -
Prompt-Free Universal Region Proposal Network
Paper • 2603.17554 • Published • 3 -
COT-FM: Cluster-wise Optimal Transport Flow Matching
Paper • 2603.13395 • Published • 2 -
A Subgoal-driven Framework for Improving Long-Horizon LLM Agents
Paper • 2603.19685 • Published • 18 -
Reasoning as Compression: Unifying Budget Forcing via the Conditional Information Bottleneck
Paper • 2603.08462 • Published • 21 -
Breaking the Capability Ceiling of LLM Post-Training by Reintroducing Markov States
Paper • 2603.19987 • Published • 9 -
EgoForge: Goal-Directed Egocentric World Simulator
Paper • 2603.20169 • Published • 9 -
UniGRPO: Unified Policy Optimization for Reasoning-Driven Visual Generation
Paper • 2603.23500 • Published • 33 -
2Xplat: Two Experts Are Better Than One Generalist
Paper • 2603.21064 • Published • 24 -
Self-Improving World Modelling with Latent Actions
Paper • 2602.06130 • Published • 31 -
InftyThink+: Effective and Efficient Infinite-Horizon Reasoning via Reinforcement Learning
Paper • 2602.06960 • Published • 14