[2602.11389] Causal-JEPA: Learning World Models through Object-Level Latent Interventions

Abstract:World models require robust relational understanding to support prediction, reasoning, and control. While object-centric representations provide a useful abstraction, they are not sufficient to capture interaction-dependent dynamics. We therefore propose C-JEPA, a simple and flexible object-centric world model that extends masked joint embedding prediction from image patches to object-centric representations. By applying object-level masking that requires an object's state to be inferred from other objects, C-JEPA induces latent interventions with counterfactual-like effects and prevents shortcut solutions, making interaction reasoning essential. Empirically, C-JEPA leads to consistent gains in visual question answering, with an absolute improvement of about 20\% in counterfactual reasoning compared to the same architecture without object-level masking. On agent control tasks, C-JEPA enables substantially more efficient planning by using only 1\% of the total latent input features required by patch-based world models, while achieving comparable performance. Finally, we provide a formal analysis demonstrating that object-level masking induces a causal inductive bias via latent interventions. Our code is available at this https URL.

View PDF HTML (experimental) Abstract:World models require robust relational understanding to support prediction, reasoning, and control. While object-centric representations provide a useful abstraction, they are not sufficient to capture interaction-dependent dynamics. We therefore propose C-JEPA, a simple and flexible object-centric world model that extends masked joint embedding prediction from image patches to object-centric representations. By masking object-level latents and requiring each masked object state to be inferred from the surrounding context, C-JEPA imposes structured…

saved by

• Henry Zhang

related reading

• The Annotated JEPA | Elements of a Vector Spaceelonlit.com
• LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixelsle-wm.github.io
• [2506.09985] V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planningarxiv.org
• [2602.10098] VLA-JEPA: Enhancing Vision-Language-Action Model with Latent World Modelarxiv.org
• V-JEPA: The next step toward advanced machine intelligenceai.meta.com
• [2603.14482] V-JEPA 2.1: Unlocking Dense Features in Video Self-Supervised Learningarxiv.org
• [2509.14252] LLM-JEPA: Large Language Models Meet Joint Embedding Predictive Architecturesarxiv.org
• pdfopenreview.net
• Stanford CS25: Transformers United V6 I From Representation Learning to World Modelingyoutube.com
• [2507.19457] GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learningarxiv.org
• The first AI model based on Yann LeCun’s vision for more human-like AIai.facebook.com
• arxiv.org/pdf/2511.08544arxiv.org