【目录】
本期的 15 篇论文如下:
[] 🔍 Your UnEmbedding Matrix is Secretly a Feature Lens for Text Embeddings(你的解嵌入矩阵实际上是文本嵌入的隐式特征透镜)
[] 🤝 SoCRATES: Towards Reliable Automated Evaluation of Proactive LLM Mediation across Domains and Socio-cognitive Variations(SoCRATES:面向跨领域和社会认知变化的主动式大语言模型调解的可靠自动评估)
[] 🎧 MMAE: A Massive Multitask Audio Editing Benchmark(MMAE:大规模多任务音频编辑基准)
[] 🧬 GENEB: Why Genomic Models Are Hard to Compare(GENEB:为什么基因组模型难以比较)
[] 🌍 AnchorWorld: Embodied Egocentric World Simulation with View-based Evolution Customization(AnchorWorld:基于视角演化定制的具身自我中心世界模拟)
[] 🎨 Direct 3D-Aware Object Insertion via Decomposed Visual Proxies(通过分解视觉代理实现直接的三维感知物体插入)
[] 🤖 Robots Need More than VLA and World Models(机器人需要的不仅仅是VLA与世界模型)
[] 🛠 When Tools Fail: Benchmarking Dynamic Replanning and Anomaly Recovery in LLM Agents(当工具失效时:大语言模型代理中的动态重规划和异常恢复基准测试)
[] 🧠 SubtleMemory: A Benchmark for Fine-Grained Relational Memory Discrimination in Long-Horizon AI Agents(SubtleMemory:面向长周期AI代理的细粒度关系记忆辨别基准)
[] 🧠 OpenSkill: Open-World Self-Evolution for LLM Agents(OpenSkill:面向LLM智能体的开放世界自我进化框架)
[] 🌍 UniSHARP: Universal Sharp Monocular View Synthesis(UniSHARP:通用锐利单目视图合成)
[] 🏃 LIMMT: Less is More for Motion Tracking(LIMMT:少即是多用于运动追踪)
[] 👁 Watch, Remember, Reason: Human-View Video Understanding with MLLMs(观看、记忆、推理:基于多模态大语言模型的人类视角视频理解)
[] 🎙 dots.tts Technical Report(dots.tts技术报告)
[] 🧠 Thinking with Imagination: Agentic Visual Spatial Reasoning with World Simulators(用想象力思考:基于世界模拟器的具身视觉空间推理)

【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
【赞助商】
OpenClaw快报
每天五分钟,听听 OpenClaw 快报,带你了解最新动态和业内讨论
传送门 www.xiaoyuzhoufm.com
