This work proposes tfour, a twin-critic method that calibrates token-level advantages from a single generated trajectory, and describes optimal mixing under the signal-retention constraint and establishes an upper bound on residual mean-induced drift.
Nan Qiao, Ye-Bin Yang, Wei-Nong Wang et al.· 0 citations
Experiments on Qwen3-VL-4B and Qwen3-VL-8B show that incorporating pattern-specific penalties into reinforcement learning can mitigate cross-mode misalignment while incurring a marginal task performance trade-off.
Xin-Ming Wang, Wei-Nong Wang, Hong-Ming Yang et al.· 2 citations· ⚡1
This report presents HyMobileAgent, a mobile GUI agent built on Hy3.0-VL-A3B, a vision-native foundation model featuring native any-resolution input, an A3B-scale deployment budget, and a 32K context window to model extended interaction histories.
H. Team, Hua-Wen Shen, Zheng-Yang Tang et al.· arXiv.org· 2 citations· ⚡1
HuyuanOCR-1.5 ranks among the top-tier end-to-end OCR solutions on OmniDocBench v1.6 while achieving new performance milestones across these long-tail tasks, and proposes Agentic Data Flow, an agent-driven data construction system that transforms model weaknesses into executable data requirements and autonomously perfo...