arXiv · cs.AI · Hao Yang, Shuoyang Liu, Xiaoyu Lin, Zhou Fang, Maji Huang, Zhaosong Huang, Ruichao Mao, Teng Guo, Hai Rao · 9 小时前模型评分39
UI-UG 2:用细粒度监督与强化学习提升移动 UI 理解与生成
UI-UG 2: Advancing Mobile UI Understanding and Generation with Fine-Grained Supervision and Reinforcement Learning
导读
UI-UG 2 是一个 4B 参数的多模态大语言模型(MLLM,Multimodal Large Language Model),在自建 80 类两级 UI 分类基准上取得所评估模型中最强的指代与密集定位结果,并在 4B 规模下取得最佳综合生成得分。
分类置信度
中 · 待人工复核
- 只有单一媒体来源,需要人工复核。
来源级别:媒体与个人
事件类型:研究发布
证据摘录
- “Inspired by the unified direction introduced by UI-UG, we present UI-UG 2, a compact 4B-parameter multimodal large language model (MLLM) that advances mobile UI Understanding and Generation with broader task coverage, a front-end-aligned taxonomy, and online render-aware reinforcement learning.”
来源:arXiv · cs.AI · arxiv.org