Current paradigms in artificial intelligence concentrate on scaling the capabilities of individual models, yet the collective behaviour of interacting agents is shaped by the topology of their interactions rather than by individual cognition alone. Here we show that the binding constraint on collective behaviour in age...
Shuo Lu, Wei-Cheng Meng, Ai-Jing Yu et al.· 1 citation
Temperature-Grouped Reinforcement Learning (TGRL), which turns temperature-induced diversity into an explicit training signal, and reaches equivalent accuracy up to 36% faster than strong RLVR baselines without expanding the rollout budget.
Zi-Han Lin, Xiao-Han Wang, Jie Cao et al.· 0 citations
Across seven benchmarks and four base VLMs, VLM-in-Sandbox achieves the highest sample-weighted average accuracy among Vanilla VLM, Append-only Sandbox, and the proposed method, and is identified as a central abstraction for sandboxed VLM agents.
He-Xiong Yang, Ming-Rui Chen, Jie Cao et al.· 0 citations
OSOL, which designates a focus domain at each iteration, uses the preceding checkpoint footprint to rank token-level rebound risk, and applies a drift-ranked, adaptively scaled correction within the standard GRPO update, is proposed, which suppresses the targeted cross-step output backtracking component.
Zi-Han Lin, Xiao-Han Wang, Jie Cao et al.· 1 citation
BASM is proposed, which augments each skill with explicit boundary fields, which transforms each retrieved skill from an unconditional action template into state-conditioned guidance: the agent applies the skill when its conditions hold, suppresses inapplicable tool calls when they do not, and issues targeted repairs w...
Zi-Han Lin, Zhen-Yu Chen, Jiawen Wei et al.· 1 citation
TREK (Teacher-Routed Exploration via Forward KL), a simple staged procedure that uses distillation not for imitation but for exploration support expansion, achieves high success rates early in training while unaided GRPO requires substantially more optimization steps to reach comparable levels.
