Quantum-JEPA: Quantum Born Priors for Latent-Action JEPA World Models
We introduce Quantum-JEPA, a latent-action JEPA framework that learns distributions of future representations using a conditional quantum Born prior.
ProofPaper ↗
Key points
- However, quantum generative models remain underexplored for this task.
- We train the quantum prior through the quality and diversity of these predictions using a sampled energy objective.
- Our quantum priors combine competitive predictive quality with a parameter-efficient sampling structure: they avoid an exponentially sized categorical output layer while supporting correlated latent actions.
- These results establish a trainable quantum component for JEPA future prediction and a basis for exploring quantum-enhanced world models at larger scales.
Sources (1)
- [1]Quantum-JEPA: Quantum Born Priors for Latent-Action JEPA World ModelsarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 20, 09:33 AM
We introduce Quantum-JEPA, a latent-action JEPA framework that learns distributions of future representations using a conditional quantum Born prior.
However, quantum generative models remain underexplored for this task.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 20, 2026EndoWorld-JEPA: Motion-Guided Global-Local Latent World Modeling for Endoscopic Video Representation Learning
- Oct 20, 2026LaLa-World: Scaling Latent World Models with Natural Language Actions
- Oct 20, 2026Slot-JEPA: Object-Structured Prediction Interfaces for Latent World Models
- Oct 19, 2026Learning Goal-Conditioned Subgoals through Frozen Hierarchical World Models
- Oct 19, 2026AnyGoal: On World Model Generalization Beyond Expert Rollouts
- Oct 19, 2026Ropedia W0: Scaling Loco-Manipulation World Models across Embodiments with Human Experience