Results demonstrate that generative video rendering mitigates the visual sim-to-real gap for zero-shot policy transfer, and shows that policy performance improves with more generated videos per simulation trajectory, increasing opening-task success by 65 percentage points.
Huang Huang, Wen-Si Ai, Zi-Yu Chen et al.· 0 citations
The Neverwhere Benchmark Suite is developed, comprised of over sixty 3D Gaussian Splatting reconstructions of urban indoor and outdoor scenes, to encourage large-scale and reproducible robot evaluation by making it easier to create and integrate Gaussian splats-based reconstructions into simulated continuous testing se...
Zi-Yu Chen, Henghui Bao, Hao-Ran Chang et al.· 0 citations
WildCity, a real-world multimodal dataset collected by autonomous fleets traversing complex urban environments, aims to catalyze progress not only in city-scale rendering, but more broadly in the pursuit of AI that can perceive, remember, and reason across space at a scale comparable to human cognition.
Xiangyu Han, Mengyu Yang, Jiaqi Li et al.· arXiv.org· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.