JEPA Policy: Diffusion-Free Imitation Learning via Paired Action and Future Rep
Overview
JEPA Policy improves imitation learning by predicting both action chunks and future state representations simultaneously within a single Transformer. It achieves an average success rate of 83.0% across nine simulation tasks, up from 77.4%, with minimal latency increase. The method avoids generating detailed future videos, instead focusing on latent representations that capture task-relevant outcomes.
From Robotics — Robot learning news discovery
Event evidence and corrections
0 attributed source owners. Ownership does not establish independent confirmation. Quantities are reported separately and are never added together.
Environment: simulation
“of success rates across 9 robot tasks For simulation evaluation, we used MokaMoka, MugMug, Square, Tool Hang, Transport, Coffee Preparation, Kitchen, Three Piece Assembly, and Hammer Cleanup. JEPA Policy outperformed action-only MIP in all 9 tasks, with the average improving by 5.6 points from 77.4% to 83.0%. The largest improvements were in Coffee Preparation, from 57.5% to 77.5%, and Tool Hang, from 65.8% to 75.0%.”
Exact source · revision 2Source owner not reported
Reported quantity · latency increase: 0.29 other · Basis not reported
“as a pair. It allows both tokens to see each other within a single Transformer and refines coarse predictions through two forward passes. Without using diffusion's iterative denoising, it increased the average success rate from 77.4% to 83.0% across nine simulation tasks while keeping the latency increase relative to MIP at 0.29ms. Title: JEPA Policy: Diffusion-Free Imitation Learning via Paired Action and Future Rep”
Exact source · revision 2Source owner not reported
Reported quantity · latency: 13.2 other · Basis not reported · Differing source assertions
“obot domains. Diagnosing future anomalies in 13.2ms Figure 6: Comparison of decision latency between JEPA Policy and Diffusion Policy Measured on the same accelerator with batch size 1 on Tool Hang, MIP was 12.9ms and JEPA Policy was 13.2ms. The additional cost of the future branch is 0.29ms, and the parameter increase is 0.79M. A 100-step Diffusion Policy takes 439.5ms, while JEPA Policy can make decisions at approx”
Exact source · revision 2Source owner not reported
Reported quantity · latency: 12.9 other · Basis not reported · Differing source assertions
“obot domains. Diagnosing future anomalies in 13.2ms Figure 6: Comparison of decision latency between JEPA Policy and Diffusion Policy Measured on the same accelerator with batch size 1 on Tool Hang, MIP was 12.9ms and JEPA Policy was 13.2ms. The additional cost of the future branch is 0.29ms, and the parameter increase is 0.79M. A 100-step Diffusion Policy takes 439.5ms, while JEPA Policy can make decisions at approx”
Exact source · revision 2Source owner not reported
Artifact availability · paper: available
“resentation Prediction URL: https://arxiv.org/abs/2609.09630 Affiliation: Anyverse Dynamics Authors: Jie Xu, Kangjin Yu, Ziyi Jin, Junjie Gao, Liqing Chen, Yixian Li, Shuai Tian, Zhongpu Xia Figure 1: Design comparison of Diffusion Policy, latent JEPA, ACT-JEPA, and JEPA Policy Background: Action labels lack the meaning of results Behavior cloning directly regresses expert actions at:t+H from observations ot. While t”
Exact source · revision 2Source owner not reported
Evidence dependency changes (2)
- 2026-10-30T21:32:29.579Z · evidence updated · source revision 2. Evidence extraction was updated; current source attributions are shown above.
- 2026-10-30T21:32:09.313Z · evidence updated · source revision 2. Evidence extraction was updated; current source attributions are shown above.
Report timeline
Follow attributed reports and material updates.
- Robotics — Robot learning news discoverySignalJEPA Policy: Diffusion-Free Imitation Learning via Paired Action and Future Representation Prediction
JEPA Policy improves imitation learning by predicting both action chunks and future state representations simultaneously within a single Transformer. It achieves an average success rate of 83.0% across nine simulation tasks, up from 77.4%, with minimal latency increase. The method avoids generating detailed future videos, instead focusing on latent representations that capture task-relevant outcomes.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).