# Sub-JEPA: Subspace Gaussian Regularization for Stable End-to-End World Models

arXiv: https://arxiv.org/abs/2605.09241 · HTML: https://mlsanity.com/paper/2605.09241/

## Key facts

- Cited by 16 papers in the map (data through 14 Oct 2026).
- Placed in Latent predictive world models, one of 11 directions in World models.
- Builds on 20 of its 37 references that are in the map.
- First posted on arXiv 10 May 2026.

## Placed in

- World models / [Latent predictive world models](https://mlsanity.com/wm/latent-predictive-world-models/)

## Builds on (most cited first, 20 of 20)

- [A Simple Framework for Contrastive Learning of Visual Representations](https://mlsanity.com/paper/2002.05709/) (2020)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Momentum Contrast for Unsupervised Visual Representation Learning](https://mlsanity.com/paper/1911.05722/) (2019)
- [Bootstrap your own latent: A new approach to self-supervised Learning](https://mlsanity.com/paper/2006.07733/) (2020)
- [Diffusion Policy: Visuomotor Policy Learning via Action Diffusion](https://mlsanity.com/paper/2303.04137/) (2023)
- [Barlow Twins: Self-Supervised Learning via Redundancy Reduction](https://mlsanity.com/paper/2103.03230/) (2021)
- [Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture](https://mlsanity.com/paper/2301.08243/) (2023)
- [DeepMind Control Suite](https://mlsanity.com/paper/1801.00690/) (2018)
- [Deep Variational Information Bottleneck](https://mlsanity.com/paper/1612.00410/) (2016)
- [Recurrent World Models Facilitate Policy Evolution](https://mlsanity.com/paper/1809.01999/) (2018)
- [Exact solutions to the nonlinear dynamics of learning in deep linear neural networks](https://mlsanity.com/paper/1312.6120/) (2013)
- [Revisiting Feature Prediction for Learning Visual Representations from Video](https://mlsanity.com/paper/2404.08471/) (2024)
- [Temporal Difference Learning for Model Predictive Control](https://mlsanity.com/paper/2203.04955/) (2022)
- [DINO-WM: World Models on Pre-trained Visual Features enable Zero-shot Planning](https://mlsanity.com/paper/2411.04983/) (2024)
- [Understanding Dimensional Collapse in Contrastive Self-supervised Learning](https://mlsanity.com/paper/2110.09348/) (2021)
- [Transformers are Sample-Efficient World Models](https://mlsanity.com/paper/2209.00588/) (2022)
- [Whitening for Self-Supervised Representation Learning](https://mlsanity.com/paper/2007.06346/) (2020)
- [LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixels](https://mlsanity.com/paper/2603.19312/) (2026)
- [OGBench: Benchmarking Offline Goal-Conditioned RL](https://mlsanity.com/paper/2410.20092/) (2024)
- [AI-Generated Video Detection via Perceptual Straightening](https://mlsanity.com/paper/2507.00583/) (2025)

## Built on by (most cited first, 16 of 16)

- [Fast LeWorldModel](https://mlsanity.com/paper/2606.26217/) (2026)
- [Delta-JEPA: Learning Action-Sensitive World Models via Latent Difference Decoding](https://mlsanity.com/paper/2606.31232/) (2026)
- [PhyLatent: Learning Dynamics-Relevant Representations for JEPA World Models](https://mlsanity.com/paper/2608.05720/) (2026)
- [QQWorld: Quantile-Quantile Matching for World Model Regularization](https://mlsanity.com/paper/2607.28415/) (2026)
- [DWM: Separating World Effects from Actions in Latent World Models](https://mlsanity.com/paper/2607.18715/) (2026)
- [ProWorld: Progress-Aware Hyperbolic World Models for Long-Horizon Visual Goal Reaching](https://mlsanity.com/paper/2608.01926/) (2026)
- [Capturing Cardiac Cyclicity through Phase-Equivariant Self-Supervised Learning](https://mlsanity.com/paper/2608.21147/) (2026)
- [FIRM-WM: State-factorized factual-interventional recurrent modeling for reward-free visual planning](https://mlsanity.com/paper/2609.22816/) (2026)
- [AD-WM: Action-Discriminative World Models for Counterfactual Model Predictive Control](https://mlsanity.com/paper/2609.30264/) (2026)
- [Adaptive Latent Capacity for World Models](https://mlsanity.com/paper/2609.32921/) (2026)
- [Hamiltonian JEPA: Action-Conditioned World Models with an Inherited Control State](https://mlsanity.com/paper/2609.33497/) (2026)
- [Beyond One-Step Accuracy: State-Affine Latent Transition for Reliable Visual Planning](https://mlsanity.com/paper/2609.33595/) (2026)
- [FlexiWorld: Learning and Planning via Flexible Action Chunks Across Multiple Time Scales](https://mlsanity.com/paper/2609.35138/) (2026)
- [Beyond a single latent space: a dual-latent world model for long-horizon planning](https://mlsanity.com/paper/2609.37644/) (2026)
- [Learning Commute-Time-Preserving World Models for Planning](https://mlsanity.com/paper/2610.01373/) (2026)
- [Frozen in a Frame: The Velocity Blind Spot in JEPA World Models](https://mlsanity.com/paper/2610.04585/) (2026)

## Abstract (arXiv)

Joint-Embedding Predictive Architectures (JEPAs) provide a simpleframework for learning world models by predicting future latent representations.However, JEPA training is subject to a bias-variance tradeoff.Without sufficient structural constraints, excessive representationalvariance causes the model to collapse to trivial solutions.The recent LeWorldModel (LeWM) shows that this issue can be alleviated bysimply constraining latent embeddings with an isotropic Gaussian prior.However, latent representations inherently lie on low-dimensional manifoldswithin a high-dimensional ambient space, and enforcing an isotropic Gaussianprior directly in this ambient space introduces an overly strong bias.In this work, we propose ame, which seeks a favorable operatingpoint on the bias-variance frontier by applying Gaussian constraints inmultiple random subspaces rather than in the originalembedding space.This design relaxes the global constraint while preserving itsanti-collapse effect, leading to a better balance between trainingstability and representation flexibility.Extensive experiments across fourcontinuous-control environments demonstrate that consistentlyoutperforms LeWM with very clear margins.Our method is simple yet effective, and serves as a strong baseline for future JEPA-based world model research.fdefinedeeemodeThe code is available at https://github.com/intcomp/Sub-JEPA.