# Spectral-Target Physical Latent Structuring for JEPA-Style World Models

arXiv: https://arxiv.org/abs/2609.04264 · HTML: https://mlsanity.com/paper/2609.04264/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Latent predictive world models, one of 11 directions in World models.
- Builds on 24 of its 35 references that are in the map.
- First posted on arXiv 2 Sep 2026.

## Placed in

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

## Builds on (most cited first, 24 of 24)

- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [End-to-End Object Detection with Transformers](https://mlsanity.com/paper/2005.12872/) (2020)
- [World Models](https://mlsanity.com/paper/1803.10122/) (2018)
- [Mastering Diverse Domains through World Models](https://mlsanity.com/paper/2301.04104/) (2023)
- [Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture](https://mlsanity.com/paper/2301.08243/) (2023)
- [SAM 3: Segment Anything with Concepts](https://mlsanity.com/paper/2511.16719/) (2025)
- [Revisiting Feature Prediction for Learning Visual Representations from Video](https://mlsanity.com/paper/2404.08471/) (2024)
- [DINO-WM: World Models on Pre-trained Visual Features enable Zero-shot Planning](https://mlsanity.com/paper/2411.04983/) (2024)
- [LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixels](https://mlsanity.com/paper/2603.19312/) (2026)
- [LeJEPA: Provable and Scalable Self-Supervised Learning Without the Heuristics](https://mlsanity.com/paper/2511.08544/) (2025)
- [Understanding and Improving Layer Normalization](https://mlsanity.com/paper/1911.07013/) (2019)
- [Learning from Reward-Free Offline Data: A Case for Planning with Latent Dynamics Models](https://mlsanity.com/paper/2502.14819/) (2025)
- [VideoCutLER: Surprisingly Simple Unsupervised Video Instance Segmentation](https://mlsanity.com/paper/2308.14710/) (2023)
- [AdaJEPA: An Adaptive Latent World Model](https://mlsanity.com/paper/2606.32026/) (2026)
- [Predictive but Not Plannable: RC-aux for Latent World Models](https://mlsanity.com/paper/2605.07278/) (2026)
- [VISReg: Variance-Invariance-Sketching Regularization for JEPA training](https://mlsanity.com/paper/2606.02572/) (2026)
- [Fast LeWorldModel](https://mlsanity.com/paper/2606.26217/) (2026)
- [Temporal-Distance JEPA: Plan-Aware Representation Learning for Latent World Model Predictive Control](https://mlsanity.com/paper/2607.25337/) (2026)
- [Is Forward Prediction Enough? Physical State Grounding for JEPA World Models](https://mlsanity.com/paper/2608.06799/) (2026)
- [Why and How Auxiliary Tasks Improve JEPA Representations](https://mlsanity.com/paper/2509.12249/) (2025)
- [QQWorld: Quantile-Quantile Matching for World Model Regularization](https://mlsanity.com/paper/2607.28415/) (2026)
- [Temporally Centered SIGReg Improves LeWorldModel Representations for Robot Policy Learning](https://mlsanity.com/paper/2607.26924/) (2026)
- [Depth-Regularized JEPA World Models Learn More Transferable Representations from Real Outdoor Robot Data](https://mlsanity.com/paper/2607.16314/) (2026)
- [Mind the Gap: Promises and Pitfalls of Hierarchical Planning in LeWorldModel](https://mlsanity.com/paper/2607.12547/) (2026)

## Abstract (arXiv)

Latent world models have become increasingly popular as a method to predict and plan in latent space rather than pixel space. Recent architectures, such as LeWorldModel (LeWM), jointly train the encoder and predictor using regularization techniques like SIGReg to prevent representation collapse. Even with such regularization preventing representation collapse, we identify a new world model failure mode of physical representation laziness, particularly noted in highly dynamic environments. For these lazy cases, the learned latent states do not collapse but nonetheless fail to represent key physical properties, causing ubiquitous downstream planning failure. To resolve this issue, we propose training-time auxiliary supervision with a lightweight "Fourier auxiliary head", which enforces physically-informed structuring of the latent space with no additional inference-time cost and can be generalized to any environment. Experimentally, we show that the auxiliary head substantially improves planning success rates in dynamic environments where the baseline LeWM exhibits physical representation laziness. It also leads to modest improvements in other environments, even when the baseline does not exhibit physical representation laziness. We further observe superior planning performance being accompanied by higher latent space correlations with key physical properties, indicating both the ability of our method to physically structure latent states and the potential planning-side benefit to the learned representation being physically structured. We also see in low-data regimes, auxiliary supervision is particularly impactful in increasing success rate. These findings support the use of our Fourier auxiliary head method to improve both overall success rate and data efficiency, while avoiding representation laziness in latent world models.