…However, existing approaches remain fundamentally limited by their static action sets and lack of principled long-horizon context management, hindering their ability to accumulate reusable experience across tasks and operate reliably in…
…Prior studies often rely on heuristic seed expansion s for data synthesis, which severely limits both novelty and difficulty. Consequently, the training value of such data fails to scale proportionally with the…
…Furthermore, statistical analysis uncovers a strong positive linear correlation between an VLM's performance on logic reasoning and long-video understanding benchmarks, establishing agentic capability scaling as a new paradigm for multimodal…
…To address these limitations, we propose BEAM (Binary Expert Activation Masking), a novel method that learns token-adaptive expert selection via trainable binary masks . With a straight-through estimator and an auxiliary…
…AI-generated summary Training multimodal large language models has long been limited by the scarcity of high-quality paired multimodal data. Recent studies show that the shared representation space of pretrained multimodal…
…Wanli Li , Bowen Zhou , , , Yifan Yang , , Abstract WeaveBench presents a comprehensive benchmark for evaluating computer-use agents across multiple interfaces, revealing significant challenges in long-horizon task orchestration and highlighting the limitations…
…AI-generated summary Large reasoning models , such as OpenAI o1 and DeepSeek-R1, tend to become increasingly verbose as their reasoning capabilities improve. These inflated Chain-of-Thought (CoT) trajectories often exceed…
…Generated by Qwen/Qwen2.5-Coder-32B-Instruct Embodied intelligence is often studied through specialized models for individual tasks such as manipulation or navigation, resulting in fragmented capabilities and limited generalization across…
…Generated by Qwen/Qwen2.5-Coder-32B-Instruct Spatial reasoning is a fundamental capability for vision-language models (VLMs) deployed in real-world environments. However, visual observations are inherently limited representations of…
…most existing methods still depend on stronger teacher models or heavily curated difficult datasets, limiting scalable capability improvement. In this paper, we introduce DenoiseRL, a reinforcement learning framework that substitutes external supervision…