2025.08.27 | 物理模型评估显不足;树算法优化提效降本

2025.08.27 | 物理模型评估显不足;树算法优化提效降本

7分钟 ·
播放数79
·
评论数0

本期的 15 篇论文如下:

00:23 🔬 CMPhysBench: A Benchmark for Evaluating Large Language Models in Condensed Matter Physics(CMPhysBench:用于评估凝聚态物理中大语言模型的基准测试)

00:57 🌳 TreePO: Bridging the Gap of Policy Optimization and Efficacy and Inference Efficiency with Heuristic Tree-based Modeling(TreePO: 通过启发式树建模弥合策略优化与效果和推理效率之间的差距)

01:21 🗣 VibeVoice Technical Report(VibeVoice技术报告)

01:45 🔨 VoxHammer: Training-Free Precise and Coherent 3D Editing in Native 3D Space(VoxHammer:在原生3D空间中无需训练的精确连贯3D编辑)

02:13 💡 Spacer: Towards Engineered Scientific Inspiration(Spacer:迈向工程化的科学灵感)

02:45 🧠 OmniHuman-1.5: Instilling an Active Mind in Avatars via Cognitive Simulation(OmniHuman-1.5:通过认知模拟为数字人注入活跃思维)

03:10 🧠 UltraMemV2: Memory Networks Scaling to 120B Parameters with Superior Long-Context Learning(UltraMemV2:扩展至1200亿参数的具有卓越长上下文学习能力的记忆网络)

03:36 ⚡ Pixie: Fast and Generalizable Supervised Learning of 3D Physics from Pixels(Pixie: 从像素中快速且可泛化的3D物理监督学习)

04:04 🎥 Autoregressive Universal Video Segmentation Model(自回归通用视频分割模型)

04:30 🎬 Wan-S2V: Audio-Driven Cinematic Video Generation(Wan-S2V:音频驱动的电影级视频生成)

04:56 🎬 CineScale: Free Lunch in High-Resolution Cinematic Visual Generation(CineScale:高分辨率电影视觉生成中的免费午餐)

05:22 🔷 FastMesh:Efficient Artistic Mesh Generation via Component Decoupling(FastMesh: 通过组件解耦实现高效艺术网格生成)

05:45 📊 ReportBench: Evaluating Deep Research Agents via Academic Survey Tasks(ReportBench:通过学术调查任务评估深度研究代理)

06:13 🧠 ThinkDial: An Open Recipe for Controlling Reasoning Effort in Large Language Models(ThinkDial:一种控制大型语言模型推理努力的开源方法)

06:42 🧠 MovieCORE: COgnitive REasoning in Movies(MovieCORE:电影中的认知推理)

【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