mlsanityData to 14 Oct 2026

Learning State-Space Models of Dynamic Systems from Arbitrary Data using Joint Embedding Predictive Architectures

arXiv 2508.10489 ↗First posted 14 Aug 2025cs.LGcited by 2

On the map: World models → Latent predictive world models

Abstract (arXiv)

With the advent of Joint Embedding Predictive Architectures (JEPAs), which appear to be more capable than reconstruction-based methods, this paper introduces a novel technique for creating world models using continuous-time dynamic systems from arbitrary observation data. The proposed method integrates sequence embeddings with neural ordinary differential equations (neural ODEs). It employs loss functions that enforce contractive embeddings and Lipschitz constants in state transitions to construct a well-organized latent state space. The approach's effectiveness is demonstrated through the generation of structured latent state-space models for a simple pendulum system using only image data. This opens up a new technique for developing more general control algorithms and estimation techniques with broad applications in robotics.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Placed in Latent predictive world models, one of 11 directions in World models. Builds on 27 of its 29 references that are in the map. First posted on arXiv 14 Aug 2025.

Builds on

27 of 29 references.

Closest ancestors · 4

Same direction, newest first

  1. 2024Learning and Leveraging World Models in Visual Representation Learning1 Mar
  2. 2023Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture19 Jan
  3. 2022Joint Embedding Predictive Architectures Focus on Slow Features20 Nov
  4. 2022Temporal Difference Learning for Model Predictive Control9 Mar

Other references · 23

  1. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 12,049
  2. 2018Representation Learning with Contrastive Predictive Codingcited by 7,777
  3. 2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,750
  4. 2018Neural Ordinary Differential Equationscited by 3,779
  5. 2020Exploring Simple Siamese Representation Learningcited by 2,866
  6. 2020Unsupervised Learning of Visual Features by Contrasting Cluster Assignmentscited by 2,857
  7. 2021Barlow Twins: Self-Supervised Learning via Redundancy Reductioncited by 1,838
  8. 2019Dream to Control: Learning Behaviors by Latent Imaginationcited by 1,653
  9. 2021VICReg: Variance-Invariance-Covariance Regularization for Self-Supervised Learningcited by 1,057
  10. 2023TD-MPC2: Scalable, Robust World Models for Continuous Controlcited by 512
  11. 2020Learning Invariant Representations for Reinforcement Learning without Reconstructioncited by 436
  12. 2021Understanding Dimensional Collapse in Contrastive Self-supervised Learningcited by 378
11 more
  1. 2015Embed to Control: A Locally Linear Latent Dynamics Model for Control from Raw Imagescited by 274
  2. 2015Action-Conditional Video Prediction using Deep Networks in Atari Gamescited by 172
  3. 2020Intriguing Properties of Contrastive Lossescited by 130
  4. 2021DreamerPro: Reconstruction-Free Model-Based Reinforcement Learning with Prototypical Representationscited by 74
  5. 2020Dreaming: Model-based Reinforcement Learning by Latent Imagination without Reconstructioncited by 66
  6. 2024Learning by Reconstruction Produces Uninformative Features For Perceptioncited by 55
  7. 2023Gradient-based Planning with World Modelscited by 19
  8. 2020Predictive Coding for Locally-Linear Controlcited by 17
  9. 2020Action-Conditional Recurrent Kalman Networks For Forward and Inverse Dynamics Learningcited by 13
  10. 2022Integrating Contrastive Learning with Dynamic Models for Reinforcement Learning from Imagescited by 12
  11. 2023Efficient Dynamics Modeling in Interactive Environments with Koopman Theorycited by 10

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026AeroJEPA: Learning Semantic Latent Representations for Scalable 3D Aerodynamic Field Modelingcited by 4
  2. 2026Data-driven Koopman mode approximation: A neural power iteration algorithmcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026LeHami: Symplectic Latent Dynamics for World Models
  2. 2026VJEPA: Variational Joint Embedding Predictive Architectures as Probabilistic World Models
  3. 2026Preserving Unstable Modes Through Inverse Dynamics in JEPA World Models
  4. 2024Deep Learning-based Approaches for State Space Models: A Selective Review
  5. 2026Semigroup-JEPA: Latent Dynamics Consistency for Zero-Shot Physics Generalization
  6. 2022Hidden Parameter Recurrent State Space Models For Changing Dynamics Scenarios
  7. 2022Continuous-time identification of dynamic state-space models by deep subspace encoding
  8. 2024Identifying Ordinary Differential Equations for Data-efficient Model-based Reinforcement Learning