This work investigates whether diffusion policies can be made more robust to viewpoint changes through different visual representations, such as 3D point clouds or pre-trained latent embeddings, such as 3D point clouds or pre-trained latent embeddings.
Aether is introduced, a simple plug-in method that applies diffusion-style random perturbations in the embedding space via controlled alpha-mixing, specifically designed to provide isotropic regularization that remains semantically consistent.
Hyesong Choi, Daeun Kim, Song Park et al.· 0 citations
This work proposes OTCLR, an optimal transport-aware framework for contrastive learning representations that generates geometry-consistent positive samples that serve as positive views that better preserve image structure while explicitly modeling spatial distributional geometry.
Akshit Nanda, Shahzad Ahmad, R. Padhy· 0 citations
This work studies what determines whether predictive supervision improves the visual representation used by a VLA policy, and finds that different prediction interfaces produce markedly different forecasts and visual representations, including in the spatial, dynamics, and action information that transfers beyond famil...
Hanseul Kim, J. Yeom, Youngjoon Jeong et al.· 0 citations
In autonomous driving, perception models often struggle to generalize to new environments due to domain shifts. While unsupervised model adaptation offers a feasible solution without labor-intensive manual labeling, existing methods that rely solely on the ego-vehicle's data often lead to inferior pseudo-labeling perfo...
Ya-Nan Ma, Yi-Hang Tao, Zheng-Ru Fang et al.· 1 citation
World models trained via pixel reconstruction can struggle in visually complex environments, where irrelevant information dominates the objective and distract the model from information relevant to planning and control. We present Contrastive World Models, an approach for learning latent dynamics models without pixel r...
This paper proposes a lightweight, plug-and-play framework that improves robustness to viewpoint shifts in Vision-Language-Action (VLA) policies without policy retraining. To our knowledge, this is the first approach to directly leverage 3D Gaussian-based novel-view synthesis for observation-space adaptation in VLA pol...
Yechan Park, Hyunji Kim· 1 citation
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
