mlsanityData to 7 Oct 2026

Value-guided action planning with JEPA world models

arXiv 2601.00844 ↗First posted 28 Dec 2025cs.LG cs.AI cs.ROcited by 19

On the map: World models → Latent predictive world models

Abstract (arXiv)

Building deep learning models that can reason about their environment requires capturing its underlying dynamics. Joint-Embedded Predictive Architectures (JEPA) provide a promising framework to model such dynamics by learning representations and predictors through a self-supervised prediction objective. However, their ability to support effective action planning remains limited. We propose an approach to enhance planning with JEPA world models by shaping their representation space so that the negative goal-conditioned value function for a reaching cost in a given environment is approximated by a distance (or quasi-distance) between state embeddings. We introduce a practical method to enforce this constraint during training and show that it leads to significantly improved planning performance compared to standard JEPA models on simple control tasks.

Key facts. Cited by 19 papers in the map (data through 7 Oct 2026). Placed in Latent predictive world models, one of 11 directions in World models. Builds on 15 of its 18 references that are in the map. First posted on arXiv 28 Dec 2025.

Builds on

15 of 18 references.

Closest ancestors · 3

Same direction, newest first

  1. 2024DINO-WM: World Models on Pre-trained Visual Features enable Zero-shot Planning7 Nov
  2. 2024Revisiting Feature Prediction for Learning Visual Representations from Video15 Feb
  3. 2023Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture19 Jan

Other references · 12

  1. 2018World Modelscited by 1,444
  2. 2020D4RL: Datasets for Deep Data-Driven Reinforcement Learningcited by 1,406
  3. 2021Offline Reinforcement Learning with Implicit Q-Learningcited by 1,245
  4. 2023HIQL: Offline Goal-Conditioned RL with Latent States as Actionscited by 133
  5. 2023Optimal Goal-Reaching Reinforcement Learning via Quasimetric Learningcited by 116
  6. 2024Foundation Policies with Hilbert Representationscited by 89
  7. 2025Learning from Reward-Free Offline Data: A Case for Planning with Latent Dynamics Modelscited by 77
  8. 2023Reinforcement Learning from Passive Data via Latent Intentionscited by 64
  9. 2022A Policy-Guided Imitation Approach for Offline Reinforcement Learningcited by 58
  10. 2022Improved Representation of Asymmetrical Distances with Interval Quasimetric Embeddingscited by 22
  11. 2019Deep Value Model Predictive Controlcited by 14
  12. 2025Infinite-Horizon Value Function Approximation for Model Predictive Controlcited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 19.

19 papers.

Sort:

By area:

  1. 2026VLA-JEPA: Enhancing Vision-Language-Action Model with Latent World Modelcited by 104
  2. 2026Predictive but Not Plannable: RC-aux for Latent World Modelscited by 16
  3. 2026VJEPA: Variational Joint Embedding Predictive Architectures as Probabilistic World Modelscited by 14
  4. 2026Human Cognition in Machines: A Unified Perspective of World Modelscited by 5
  5. 2026A Generalization Theory for JEPA-Based World Modelscited by 5
  6. 2026Temporal-Distance JEPA: Plan-Aware Representation Learning for Latent World Model Predictive Controlcited by 5
  7. 2026Qantara: Bridge-Flow Training for Multi-Paradigm JEPA Controlcited by 4
  8. 2026SCALE: State-Calibrated Latent Embeddings for JEPA Planning in the Right Geometrycited by 4
  9. 2026BiJEPA: Bi-directional Joint Embedding Predictive Architecture for Symmetric Representation Learningcited by 3
  10. 2026Latent State Design for World Models under Sufficiency Constraintscited by 3
  11. 2026AeroJEPA: Learning Semantic Latent Representations for Scalable 3D Aerodynamic Field Modelingcited by 2
  12. 2026Traj-LeWM: Path-Aware World-Model Planning via Latent Trajectory Costcited by 2
7 more
  1. 2026Grounded World Model for Semantically Generalizable Planningcited by 1
  2. 2026Intrinsic-Energy Joint Embedding Predictive Architectures Induce Quasimetric Spacescited by 0
  3. 2026Finite-Sample Metric Non-Collapse for Geometrically Supervised Latent World Models in Controlcited by 0
  4. 2026World Simulator: Queer Erotica and the Absurdity of AI Video Models That Promise the Worldcited by 0
  5. 2026Neither Adversarial Training Nor Purification: Emergent Adversarial Robustness from Oscillatory Predictive Learningcited by 0
  6. 2026I Act Therefore I Am: When Is JEPA's Action-Conditioning Enough to Learn Causal Mechanisms?cited by 0
  7. 2026Control-Geometry Straightening for Sampling-Based Latent Planningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025What Drives Success in Physical Planning with Joint-Embedding Predictive World Models?
  2. 2026FF-JEPA: Long-Horizon Planning in World Models with Latent Planners
  3. 2026Toward Physically Grounded JEPA World Models for Goal-Conditioned Robotic Planning
  4. 2026MA-JEPA: Joint-Embedding World Models for Multi-Agent Reinforcement Learning
  5. 2025JEPA for RL: Investigating Joint-Embedding Predictive Architectures for Reinforcement Learning
  6. 2026AD-E2E-JEPA: A Joint-Embedding Predictive Architecture For End-to-End Autonomous Driving
  7. 2026Hamiltonian JEPA: Action-Conditioned World Models with an Inherited Control State
  8. 2026A Lightweight Library for Energy-Based Joint-Embedding Predictive Architectures