Abstract
Robot foundation models provide strong visuomotor control, yet their performance can degrade when object positions or task instructions change. Further improvements often require post-training on substantial robot data, which can be costly to collect through methods such as teleoperation. Agentic harnesses can adapt around the model, but current self-evolving harnesses use robot trials inefficiently when deciding which code and skill changes to pursue. We introduce EmbodiedRSI, a self-evolving agentic harness that autonomously decides where to explore next and turns the resulting physical interaction into improved code and skills. EmbodiedRSI realizes this through a Fast-Slow Dual-System Architecture, in which competing code and skill hypotheses are maintained in a Hypothesis Graph. Value-of-Information Experiment Selection chooses physical experiments that can distinguish these hypotheses. Their outcomes guide Code-Skill Co-Evolution. The Slow System builds Hierarchical Memory, and Reward-Grounded Memory Learning selects effective memory according to their value for later Fast-System improvement. On RoboCasa365, EmbodiedRSI reaches 77.0% overall success and 71.3% on Composite-Unseen, compared with 40.1% for the best baseline. EmbodiedRSI also reaches 86.8% overall success on LIBERO-Pro. Beyond benchmark performance, EmbodiedRSI transfers zero-shot to real-world robot, achieving 71.3% overall success across multiple challenging tasks.
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Song, P., Liang, Z., Fu, K., Wang, M., Yang, J., & Liu, S. (2026). EmbodiedRSI: Active Continual Robot Learning Through Hypothesis-Guided Co-Evolution. https://omanscience.com/en/articles/embodiedrsi-active-continual-robot-learning-through-hypothesis-guided-co-evolution
MLA 9
Song, Python, et al. "EmbodiedRSI: Active Continual Robot Learning Through Hypothesis-Guided Co-Evolution." https://omanscience.com/en/articles/embodiedrsi-active-continual-robot-learning-through-hypothesis-guided-co-evolution.
Chicago (author–date)
Song, Python, Zhixuan Liang, Kelsey Fu, Mengdi Wang, Junfeng Yang, and Shilong Liu. 2026. "EmbodiedRSI: Active Continual Robot Learning Through Hypothesis-Guided Co-Evolution." https://omanscience.com/en/articles/embodiedrsi-active-continual-robot-learning-through-hypothesis-guided-co-evolution.
Harvard
Song, P., Liang, Z., Fu, K., Wang, M., Yang, J. and Liu, S. (2026) 'EmbodiedRSI: Active Continual Robot Learning Through Hypothesis-Guided Co-Evolution', Available at: https://omanscience.com/en/articles/embodiedrsi-active-continual-robot-learning-through-hypothesis-guided-co-evolution.
Vancouver
Song P, Liang Z, Fu K, Wang M, Yang J, Liu S. EmbodiedRSI: Active Continual Robot Learning Through Hypothesis-Guided Co-Evolution. https://omanscience.com/en/articles/embodiedrsi-active-continual-robot-learning-through-hypothesis-guided-co-evolution
IEEE
P. Song, Z. Liang, K. Fu, M. Wang, J. Yang, and S. Liu, "EmbodiedRSI: Active Continual Robot Learning Through Hypothesis-Guided Co-Evolution," https://omanscience.com/en/articles/embodiedrsi-active-continual-robot-learning-through-hypothesis-guided-co-evolution.