滤波修正扩散采样，单相机重建人景

从271篇论文中选出8篇

重点关注

DiFA: Inference-Time Forward-Process Alignment for Diffusion Models score 8
入选 HF Daily Papers；HF 热度: 5 upvotes (+2)；顶会接收: ICML
LLM-as-a-Coach: Experiential Learning for Non-Verifiable Tasks score 7
入选 HF Daily Papers；HF 热度: 11 upvotes (+3)；关键词(2): distillation, post-training

也值得关注

TraversRL: Traversable Pedestrian Pathway Generation With Reinforcement Learning score 4
关键词(1): edge；顶会接收: ECCV
Reinforcement Learning: From Algorithms To Foundation Models score 4
机构: Princeton；关键词(1): reasoning
A Geometric Perspective on Stabilizing Value Conflict Resolution score 4
关键词(2): RLHF, reasoning；顶会接收: ICML
Adaptive Mamba Neural Operators score 4
关键词(1): mamba；顶会接收: ICLR
O-VAD: Industrial Video Anomaly Detection through Object-Centric Tracking and Reasoning score 4
关键词(2): agentic, reasoning；顶会接收: ECCV
ReViV: Reconstructing the Viewer and the View in 4D from Monocular Egocentric Video score 9
入选 HF Daily Papers；HF 热度: 2 upvotes (+1)；有代码实现；顶会接收: ECCV