部署后学习拟合出R²=0.998

重点关注

• Vision Pretraining for Dense Spatial Perception score 10
  入选 HF Daily Papers；HF 热度: 35 upvotes (+4)；有代码实现；关键词(3): scaling, pretraining, embodied
• Deform360: A Massive Multi-view Visuotactile Dataset for Deformable World Models score 10
  入选 HF Daily Papers；HF 热度: 6 upvotes (+2)；有代码实现；顶会接收: ECCV
• PixWorld: Unifying 3D Scene Generation and Reconstruction in Pixel Space score 9
  入选 HF Daily Papers；HF 热度: 52 upvotes (+4)；有代码实现
• Do All Visual Tokens Matter Equally? Object-Evidence Preserving Token Merging for Vision-Language Retrieval score 9
  入选 HF Daily Papers；HF 热度: 17 upvotes (+3)；有代码实现；关键词(3): compression, pruning, vision-language
• Unified Audio Intelligence Without Regressing on Text Intelligence score 7
  入选 HF Daily Papers；HF 热度: 10 upvotes (+3)；关键词(4): distillation, MoE, agentic, reasoning
• MV-Forcing: Long Multi-View Video Generation via 4D-Grounded Spatio-Temporal Self-Forcing score 9
  入选 HF Daily Papers；HF 热度: 8 upvotes (+2)；关键词(1): distillation；顶会接收: ECCV
• Multiplayer Interactive World Models with Representation Autoencoders score 8
  入选 HF Daily Papers；HF 热度: 15 upvotes (+3)；有代码实现
• LLM-as-a-Verifier: A General-Purpose Verification Framework score 9
  入选 HF Daily Papers；HF 热度: 10 upvotes (+3)；有代码实现；关键词(7): scaling, GRPO, post-training, pre-training, agentic
• KVpop -- Key-Value Cache Compression with Predictive Online Pruning score 7
  入选 HF Daily Papers；HF 热度: 17 upvotes (+3)；关键词(3): compression, pruning, reasoning
• EdgeBench: Unveiling Scaling Laws of Learning from Real-World Environments score 7
  入选 HF Daily Papers；HF 热度: 13 upvotes (+3)；关键词(3): scaling, deployment, pretraining

也值得关注

• PixelPilot: Scalable Vision-Language-Action Models for End-to-End Autonomous Driving score 4
  关键词(3): GRPO, reasoning, vision-language；顶会接收: ECCV
• RADIANCE: Relative Adaptive Denoising with IP-Adapter for Novel Concept Enhancement score 4
  关键词(3): latency, throughput, text-to-image；顶会接收: ECCV
• Open Problems in AI Incident Governance score 4
  关键词(1): deployment；顶会接收: ICML
• Selective Disclosure Watermarking for Large Language Models score 4
  关键词(1): latency；顶会接收: ICML
• Geometric Reciprocity: Unlocking Self-Supervision for Stereoscopic Video Generation score 4
  关键词(1): synthetic data；顶会接收: ICML
• InFlux++: Real and Synthetic Data for Estimating Dynamic Camera Intrinsics score 4
  关键词(3): finetuning, leaderboard, synthetic data；顶会接收: ECCV
• Integrated Forward-Inverse Network for Lensless Image Reconstruction score 3
  顶会接收: ECCV
• Learning Structured Visual Compositional Representations for Weakly Supervised Referring Expression Comprehension score 3
  顶会接收: ECCV
• From Open Loop to Closed Loop: A Test-Time Iterative Optimization Framework for Reference-Consistent Image Generation score 3
  顶会接收: ECCV
• Learning Probabilistic Prompt for Continual Learning score 3
  顶会接收: ECCV
• MemPose: Category-level Object Pose Estimation with Memory score 3
  顶会接收: ECCV
• Fully Rotation-Equivariant Spectral-Spatial Learning for Multispectral Object Detection score 3
  顶会接收: ECCV