HuggingFace 每日AI论文速递

2025.07.22 | MiroMind-M1提升数学推理;GUI-G$^2$高斯奖励助GUI定位。

本期的 15 篇论文如下:

[00:25] 🧮 MiroMind-M1: An Open-Source Advancement in Mathematical Reasoning via Context-Aware Multi-Stage Policy Optimization(MiroMind-M1:通过上下文感知多阶段策略优化实现数学推理的开源进展)

[01:00] 🎯 GUI-G$^2$: Gaussian Reward Modeling for GUI Grounding(GUI-G$^2$: 用于GUI定位的高斯奖励建模)

[01:42] ⛓ The Invisible Leash: Why RLVR May Not Escape Its Origin(隐形束缚:RLVR为何难以摆脱其起源)

[02:53] 🏗 WebShaper: Agentically Data Synthesizing via Information-Seeking Formalization(WebShaper:通过信息寻求形式化实现代理式数据合成)

[03:20] 🤖 NoHumansRequired: Autonomous High-Quality Image Editing Triplet Mining(无需人工:自主高质量图像编辑三元组挖掘)

[04:23] 🛠 Robust 3D-Masked Part-level Editing in 3D Gaussian Splatting with Regularized Score Distillation Sampling(鲁棒的3D遮罩部件级编辑:基于正则化得分蒸馏采样的3D高斯泼溅)

[05:15] 🧠 SeC: Advancing Complex Video Object Segmentation via Progressive Concept Construction(SeC:通过渐进式概念构建推进复杂视频对象分割)

[06:19] 🤖 GR-3 Technical Report(GR-3技术报告)

[07:08] 🤖 Being-H0: Vision-Language-Action Pretraining from Large-Scale Human Videos(Being-H0:基于大规模人类视频的视觉-语言-动作预训练)

[08:12] 💡 Stabilizing Knowledge, Promoting Reasoning: Dual-Token Constraints for RLVR(稳定知识,促进推理:RLVR的双令牌约束)

[09:12] 🧠 Towards Video Thinking Test: A Holistic Benchmark for Advanced Video Reasoning and Understanding(迈向视频思维测试:一个用于高级视频推理和理解的综合基准)

[09:52] 📉 Inverse Scaling in Test-Time Compute(测试时计算中的逆向扩展)

[10:32] 💡 Gaussian Splatting with Discretized SDF for Relightable Assets(基于离散化SDF的高斯泼溅技术,用于可重光照资产)

[11:24] 🧠 STITCH: Simultaneous Thinking and Talking with Chunked Reasoning for Spoken Language Models(STITCH:口语语言模型中基于分块推理的同步思考与表达)

[12:13] ⏩ Streaming 4D Visual Geometry Transformer(流式4D视觉几何Transformer)

【关注我们】

您还可以在以下平台找到我们,获得播客内容以外更多信息

小红书: AI速递