Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Daniel, Tal, Qi, Carl, Haramati, Dan, Zadeh, Amir, Li, Chuan, Tamar, Aviv, Pathak, Deepak, Held, David
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2603.04553
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911485960126464
author Daniel, Tal
Qi, Carl
Haramati, Dan
Zadeh, Amir
Li, Chuan
Tamar, Aviv
Pathak, Deepak
Held, David
author_facet Daniel, Tal
Qi, Carl
Haramati, Dan
Zadeh, Amir
Li, Chuan
Tamar, Aviv
Pathak, Deepak
Held, David
contents 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
format Preprint
id arxiv_https___arxiv_org_abs_2603_04553
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Latent Particle World Models: Self-supervised Object-centric Stochastic Dynamics Modeling
Daniel, Tal
Qi, Carl
Haramati, Dan
Zadeh, Amir
Li, Chuan
Tamar, Aviv
Pathak, Deepak
Held, David
Machine Learning
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
title Latent Particle World Models: Self-supervised Object-centric Stochastic Dynamics Modeling
topic Machine Learning
url https://arxiv.org/abs/2603.04553