
时长:
7分钟
播放:
62
发布:
2周前
主播...
简介...
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
[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
评价...
空空如也
小宇宙热门评论...
暂无小宇宙热门评论