{"ID":6023753,"CreatedAt":"2026-07-08T01:00:23.257252134Z","UpdatedAt":"2026-07-09T07:04:12.978444665Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.05468","arxiv_id":"2607.05468","title":"Learning 4D Geometric Priors for Inference-Efficient World Action Models","abstract":"World Action Models (WAMs) have shown strong potential for robotic manipulation by jointly modeling visual future dynamics and executable action sequences. However, existing video-action co-training methods primarily optimize appearance-oriented video latents, which may insufficiently capture the temporally evolving geometry required for precise manipulation. We propose MECo-WAM, a Multi-Expert Co-Training World Action Model that injects action-relevant 4D geometric priors into video-action representations while preserving the original lightweight inference graph. During training, MECo-WAM combines video and action experts with a lightweight 4D expert supervised by relational targets from a frozen VGGT encoder. Asymmetric expert visibility prevents non-causal shortcuts from auxiliary geometry to action generation. To transfer geometric knowledge into the deployed video-action pathway, we introduce decayed 4D read-mask attention, which provides restricted current-frame geometric guidance early in training and progressively removes this dependency. We further propose action-aware temporal geometric distillation, which aligns within-frame geometric relations and their temporal evolution while emphasizing visual regions most relevant to robot actions. At deployment, all auxiliary 4D components are removed. Experiments on LIBERO (98.2%), RoboTwin 2.0 (92.6%), and challenging real-world manipulation tasks show that MECo-WAM improves manipulation performance without increasing inference cost.","short_abstract":"World Action Models (WAMs) have shown strong potential for robotic manipulation by jointly modeling visual future dynamics and executable action sequences. However, existing video-action co-training methods primarily optimize appearance-oriented video latents, which may insufficiently capture the temporally evolving ge...","url_abs":"https://arxiv.org/abs/2607.05468","url_pdf":"https://arxiv.org/pdf/2607.05468v1","authors":"[\"Jianjun Zhang\",\"Jian Zhu\",\"Taiyi Su\",\"Chong Ma\",\"Zitai Huang\",\"Yi Xu\",\"Hanli Wang\"]","published":"2026-07-06T07:10:22Z","proceeding":"cs.RO","tasks":"[\"cs.RO\",\"cs.AI\"]","methods":"[]","has_code":false}