# JEDI: Joint-Embedding Demonstration Compression for In-Context Imitation Learning

**Authors:** Jiazhao Liang, Zhewen He, Shuaihang Yuan, Hao Huang, Yi Fang  
**Lab:** New York University Abu Dhabi · **Published:** 29 Oct 2026  
**Links:** [arXiv](https://arxiv.org/abs/2610.40356) · [PDF](https://arxiv.org/pdf/2610.40356) · [This page](https://omen-bit.github.io/ai-paper-daily/paper/2610.40356.html)

## Abstract

In-context imitation learning (ICIL) adapts robot policies through test-time demonstrations, but long multimodal trajectories constrain how many examples fit in context. We ask what a compressed demonstration should preserve. We propose JEDI, which compresses demonstrations into latent slots trained to predict the query episode's next-stage keyframe in a frozen video-JEPA space. Task and process counterfactuals help the slots preserve execution details, while the predictor selects demonstrations from a bank. Post-training with keypoint tokens enables human-hand conditioning on robots. Experiments on LIBERO and real robots show that JEDI supports many-shot context; LIBERO objective controls show improved success and demonstration following over action-only compression. These findings support query-future prediction as a training objective for preserving execution-relevant information in compressed demonstrations.

---

Notes written by AI from the paper, numbers copied from the abstract. Check the paper before you rely on them. From AI Papers Daily.