Welcome to check out the #CoRL paper co-authored with Chen Feng Yue Wang : Multi Robot Scene Completion: Towards Task-agnostic Collaborative Perception, which enables co-perception with a shared spatiotemporal autoencoder (STAR). Code: github.com/coperception/s…

We've released our paper "Generating 3D-Consistent Videos from Unposed Internet Photos"! Video models like Luma generate pretty videos, but sometimes struggle with 3D consistency. We can do better by scaling them with 3D-aware objectives. 1/N page: genechou.com/kfcw

#NeurIPS2024🌟 Humans naturally retain memories of permanent elements, while ephemeral moments often slip through the cracks of memory. We endow the robot with a similar memory mechanism to only remember the environment while forgetting objects, without any human supervision.

We find training unified multimodal understanding and generation models is so easy, you do not need to tune MLLMs at all. MLLM's knowledge/reasoning/in-context learning can be transferred from multimodal understanding (text output) to generation (pixel output) even it is FROZEN!