# No Gaussian Required: Contrastive Inverse Dynamics for JEPA World Models

arXiv: https://arxiv.org/abs/2608.17542 · HTML: https://mlsanity.com/paper/2608.17542/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in Latent predictive world models, one of 11 directions in World models.
- Builds on 19 of its 24 references that are in the map.
- First posted on arXiv 18 Aug 2026.

## Placed in

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

## Builds on (most cited first, 19 of 19)

- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [Bootstrap your own latent: A new approach to self-supervised Learning](https://mlsanity.com/paper/2006.07733/) (2020)
- [Dream to Control: Learning Behaviors by Latent Imagination](https://mlsanity.com/paper/1912.01603/) (2019)
- [Curiosity-driven Exploration by Self-supervised Prediction](https://mlsanity.com/paper/1705.05363/) (2017)
- [VICReg: Variance-Invariance-Covariance Regularization for Self-Supervised Learning](https://mlsanity.com/paper/2105.04906/) (2021)
- [Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture](https://mlsanity.com/paper/2301.08243/) (2023)
- [Recurrent World Models Facilitate Policy Evolution](https://mlsanity.com/paper/1809.01999/) (2018)
- [Revisiting Feature Prediction for Learning Visual Representations from Video](https://mlsanity.com/paper/2404.08471/) (2024)
- [TD-MPC2: Scalable, Robust World Models for Continuous Control](https://mlsanity.com/paper/2310.16828/) (2023)
- [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)
- [Learning to Poke by Poking: Experiential Learning of Intuitive Physics](https://mlsanity.com/paper/1606.07419/) (2016)
- [Learning from Reward-Free Offline Data: A Case for Planning with Latent Dynamics Models](https://mlsanity.com/paper/2502.14819/) (2025)
- [Loss is its own Reward: Self-Supervision for Reinforcement Learning](https://mlsanity.com/paper/1612.07307/) (2016)
- [TACO: Temporal Latent Action-Driven Contrastive Loss for Visual Reinforcement Learning](https://mlsanity.com/paper/2306.13229/) (2023)
- [Guaranteed Discovery of Control-Endogenous Latent States with Multi-Step Inverse Models](https://mlsanity.com/paper/2207.08229/) (2022)
- [Sensorimotor World Models: Perception for Action via Inverse Dynamics](https://mlsanity.com/paper/2606.20104/) (2026)
- [CLOUD: Contrastive Learning of Unsupervised Dynamics](https://mlsanity.com/paper/2010.12488/) (2020)

## Built on by (most cited first, 5 of 5)

- [Algebraic Consistency Alone Does Not Certify Temporal Structure in Latent Action Models](https://mlsanity.com/paper/2609.23478/) (2026)
- [Representation World Model: Learning States, Transition and Executable Plans in Representation](https://mlsanity.com/paper/2609.29171/) (2026)
- [SLIP-VLA: Single-Step Latent Imagination for Policy Learning in Vision-Language-Action Models](https://mlsanity.com/paper/2609.33575/) (2026)
- [CF-JEPA: Improving Robustness of JEPA World Models via Controllability Factorization](https://mlsanity.com/paper/2610.00727/) (2026)
- [Keeping JEPA World Models Plannable When Little of the Frame Moves](https://mlsanity.com/paper/2610.03137/) (2026)

## Abstract (arXiv)

Joint-Embedding Predictive Architectures (JEPAs) learn world models by predicting future embeddings, but the objective admits a trivial solution of a constant encoder, so every practical system adds an anti-collapse mechanism (LeCun, 2022; Assran et al., 2023; Bardes et al., 2022; 2024). LeWorldModel (LeWM) prevents collapse with SIGReg, a regularizer that forces the latent distribution to match an isotropic Gaussian: the representation is stabilized by prescribing what it must look like, independently of the environment it models. We argue that the anti-collapse pressure can instead come from the transition data itself. Action-Contrastive Masked Transition Modeling (AC-MTM) keeps LeWM's forward latent-prediction objective and adds a training-only inverse-dynamics head trained with Action-NCE: each latent transition must identify the action that produced it among the other actions in the batch, a discrimination task that a collapsed encoder provably fails. The inverse branch is discarded after training, leaving test-time encoding, forward prediction, planning, and compute identical to LeWM. On four standard pixel-control tasks under a matched planning protocol, AC-MTM trains stably from scratch and matches SIGReg on average. On the harder multi-object OGBench Visual Scene task, results are consistent with the prescribed geometry becoming a bottleneck: AC-MTM reaches 80.0±2.0% success versus 58.0±2.0% for SIGReg, improving by 20-24 points in each training seed. A single 50-episode random-policy run gives a 52% baseline estimate. Contrastive inverse dynamics thus provides a distribution-free anti-collapse signal that requires no target network, stop-gradient, pretrained encoder, or reconstruction objective, and we characterize the action-space and observability assumptions under which it holds. We make our code available at https://github.com/jackboyla/action-contrastive-jepa