13 papers
HarnessEval-W: Agentifying the Evaluation of Visual Worlds
Weiliang Chen, Haowen Sun, Jun Gao +40
A benchmark should deliver more than a scalar score: what makes an evaluation trustworthy is the reasoning that justifies the score. This is especially critical for world models, w…
WAM-TTT: Steering World-Action Models by Watching Human Play at Test Time
Yusen Feng, Bingchen Han, Jiangran Lyu +13
Steering robot foundation models (RFMs) toward new task variants or user-preferred behaviors remains challenging, often requiring additional robot demonstrations, task-specific fin…
Cosmos 3: Omnimodal World Models for Physical AI
NVIDIA, :, Aditi +293
We introduce Cosmos 3, a family of omnimodal world models designed to jointly process and generate language, image, video, audio, and action sequences within a unified mixture-of-t…
KPGrasp: Scalable Keypoint Flow Matching for Dexterous Grasp Generation
Yuansen Huang, Jiayi Chen, Haoran Liu +6
Generating high-quality dexterous grasps remains challenging for learning-based methods, which often depend on carefully tuned contact losses or costly contact-based test-time refi…
LDA-1B: Scaling Latent Dynamics Action Model via Universal Embodied Data Ingestion
Jiangran Lyu, Kai Liu, Xuheng Zhang +20
Recent robot foundation models largely rely on large-scale behavior cloning, which imitates expert actions but discards transferable dynamics knowledge embedded in heterogeneous em…
Emerging Extrinsic Dexterity in Cluttered Scenes via Dynamics-aware Policy Learning
Yixin Zheng, Jiangran Lyu, Yifan Zhang +8
Extrinsic dexterity leverages environmental contact to overcome the limitations of prehensile manipulation. However, achieving such dexterity in cluttered scenes remains challengin…