Latent Particle World Models: Self-supervised Object-centric Stochastic Dynamics Modeling

Tal Daniel, Carl Qi, Dan Haramati, Amir Zadeh, Chuan Li, Aviv Tamar, Deepak Pathak, David Held

Keywords: World Model, Self-supervised, unsupervised, object-centric, video prediciton, video generation, imitation learning, latent particles, vae

TL;DR: a self-supervised object-centric world model that learns keypoints, and masks directly from videos, supports multi-modal conditioning, scaled to real-world multi-object datasets

Abstract: We introduce Latent Particle World Model (LPWM), a self-supervised object-centric world model scaled to real-world multi-object datasets and applicable in decision-making. LPWM autonomously discovers keypoints, bounding boxes, and object masks directly from video data, enabling it to learn rich scene decompositions without supervision. Our architecture is trained end-to-end purely from videos and supports flexible conditioning on actions, language, and image goals. LPWM models stochastic particle dynamics via a novel latent action module and achieves state-of-the-art results on diverse real-world and synthetic datasets. Beyond stochastic video modeling, LPWM is readily applicable to decision-making, including goal-conditioned imitation learning, as we demonstrate in the paper. Code, data, pre-trained models and video rollouts are available: https://taldatech.github.io/lpwm-web

Primary Area: unsupervised, self-supervised, semi-supervised, and supervised representation learning

Submission Number: 9321