Latent Dynamics · 2026-10-21
Quantum-JEPA: Quantum Born Priors for Latent-Action JEPA World Models
Zhiyuan Liu, Sicheng Liu, Haiyu Wu
AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.
TL;DR
Quantum-JEPA is a latent-action JEPA framework that learns a distribution of future representations using a conditional quantum Born prior. Circuit measurements generate binary latent actions that a shared predictor maps to alternative future representations. The quantum prior is trained via a sampled energy objective using finite-shot parameter-shift gradients, and the authors also test feasibility on a real QPU.
Why it matters
The authors target uncertainty in world models by learning multiple plausible futures from a single observation. They study whether sampled binary latent-action priors can produce valid, diverse alternatives and coordinated changes (including multi-agent futures), while keeping the generative structure parameter-efficient and feasible for quantum hardware inference.
Method
- Model: shared encoder + shared predictor Pψ; a conditional quantum Born prior pθ(z|h) outputs binary latent actions z via circuit measurements, then maps (h,z) to future representations to be scored against encoded targets.
- Training: sampled predictive energy loss encourages both prediction validity (closeness to targets) and diversity among sampled predictions; VISReg-style feature regularization is used.
- Quantum learning: measurement-based quantum prior updates use finite-shot parameter-shift gradients (with readout noise in reference runs), and they deploy trained priors for inference on real QPU hardware.
Abstract (from arXiv)
A single observation can admit several plausible futures, making uncertainty an essential part of world modelling. However, quantum generative models remain underexplored for this task. We introduce Quantum-JEPA, a latent-action JEPA framework that learns distributions of future representations using a conditional quantum Born prior. Circuit measurements generate binary latent actions, which a shared predictor maps to alternative future representations. We train the quantum prior through the quality and diversity of these predictions using a sampled energy objective. Our model learns distinct continuations on constructed multimodal tasks and captures latent-action correlations that support coordinated multi-agent futures. On D3IL, a benchmark of diverse human demonstrations, our trained models generate multiple obstacle-crossing routes from a common starting point without observed motion. On the controlled tasks, our classical and quantum prior implementations improve prediction validity and distributional fidelity over the evaluated variational baselines. 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. We further validate finite-shot parameter-shift training, including training with readout noise. Unlike most prior work, which focuses on simulation, we also verify the feasibility of inference with trained priors on a real QPU. These results establish a trainable quantum component for JEPA future prediction and a basis for exploring quantum-enhanced world models at larger scales.
Related papers
- LeHami: Symplectic Latent Dynamics for World Models
- LRC-JEPA: Disentangling Dynamics and Residual Context for Efficient World Models
- Exploring Temporal Context-Aware Physical Grounding in Latent World Models
- Hamiltonian JEPA: Action-Conditioned World Models with an Inherited Control State
- MDN-VJEPA: Multimodal Probabilistic Joint-Embedding World Models
Publication record
Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.
First public release: 2026-10-20 · Indexed by Pub-AI: 2026-10-21 UTC
Quantum-JEPA: Quantum Born Priors for Latent-Action JEPA World Models
Preprint posted: 2026-10-20 · Original record ↗
- arxiv
- 2610.27921
Explicit source versions (1)
- v1 · 2026-10-20
- Preprint available
Source date: 2026-10-20 · First detected: 2026-10-25 UTC · Historical baseline
Source evidence (3)
- Official evidence ↗ · OAI arXiv created; not header datestamp (official metadata)
- Official evidence ↗ · atom:published (official metadata)
- Official evidence ↗ · atom:published (official metadata)
Abstract · arxiv
A single observation can admit several plausible futures, making uncertainty an essential part of world modelling. However, quantum generative models remain underexplored for this task. We introduce Quantum-JEPA, a latent-action JEPA framework that learns distributions of future representations using a conditional quantum Born prior. Circuit measurements generate binary latent actions, which a shared predictor maps to alternative future representations. We train the quantum prior through the quality and diversity of these predictions using a sampled energy objective. Our model learns distinct continuations on constructed multimodal tasks and captures latent-action correlations that support coordinated multi-agent futures. On D3IL, a benchmark of diverse human demonstrations, our trained models generate multiple obstacle-crossing routes from a common starting point without observed motion. On the controlled tasks, our classical and quantum prior implementations improve prediction validity and distributional fidelity over the evaluated variational baselines. 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. We further validate finite-shot parameter-shift training, including training with readout noise. Unlike most prior work, which focuses on simulation, we also verify the feasibility of inference with trained priors on a real QPU. These results establish a trainable quantum component for JEPA future prediction and a basis for exploring quantum-enhanced world models at larger scales.
