# StaMo: Unsupervised Learning of Generalizable Robot Motion from Compact State Representation

Source: [StaMo: Unsupervised Learning of Generalizable Robot Motion from Compact State Representation](<https://www.sota2.com/research/paper/68fa7be21e209131cea8b091>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-15T05:00:11.244Z

Authors: Mingyu Liu, Jiuhe Shu, Hui Chen, Zeju Li, Canyu Zhao, Jiange Yang, Shenyuan Gao, Hao Chen, Chunhua Shen

Published: 2025-10-06T17:37:24.000Z

## Abstract

A fundamental challenge in embodied intelligence is developing expressive and compact state representations for efficient world modeling and decision making. However, existing methods often fail to achieve this balance, yielding representations that are either overly redundant or lacking in task-critical information. We propose an unsupervised approach that learns a highly compressed two-token state representation using a lightweight encoder and a pre-trained Diffusion Transformer (DiT) decoder, capitalizing on its strong generative prior. Our representation is efficient, interpretable, and integrates seamlessly into existing VLA-based models, improving performance by 14.3% on LIBERO and 30% in real-world task success with minimal inference overhead. More importantly, we find that the difference between these tokens, obtained via latent interpolation, naturally serves as a highly effective latent action, which can be further decoded into executable robot actions. This emergent capability reveals that our representation captures structured dynamics without explicit supervision. We name our method StaMo for its ability to learn generalizable robotic Motion from compact State representation, which is encoded from static images, challenging the prevalent dependence to learning latent action on complex architectures and video data. The resulting latent actions also enhance policy co-training, outperforming prior methods by 10.4% with improved interpretability. Moreover, our approach scales effectively across diverse data sources, including real-world robot data, simulation, and human egocentric video.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Robotic Manipulation on Real World Robot Setting](<https://www.sota2.com/research/sota/robotic-manipulation-on-real-world-robot-setting.md>): Robotic Manipulation; Real World Robot Setting; Pick Up Success Rate: 70.

- [Robot Manipulation on ManiSkill zero-shot](<https://www.sota2.com/research/sota/robot-manipulation-on-maniskill-zero-shot.md>): Robot Manipulation; ManiSkill zero-shot; Pick Cube Success Rate: 41.

- [Robotic Manipulation on LIBERO](<https://www.sota2.com/research/sota/robotic-manipulation-on-libero.md>): Robotic Manipulation; LIBERO; Spatial Success Rate: 96.8.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2510.05057>)

[PDF](<https://arxiv.org/pdf/2510.05057.pdf>)
