Paper page - ENPIRE: Agentic Robot Policy Self-Improvement in the Real World

Source: https://huggingface.co/papers/2606.19980 Authors:

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

Abstract

ENPIRE framework enables autonomous robotics research through a closed-loop system that automates policy improvement via environment feedback, policy refinement, and evolutionary code optimization.

Achieving dexterous robotic manipulation in the real world heavily relies on human supervision and algorithm engineering, which becomes a central bottleneck in the pursuit of general physical intelligence. Although emergingcoding agentscan generate code to automatealgorithm search, their successes remain largely confined in digital environments. We conjecture that the missing abstraction to automate robotics research is a repeatable feedback loop for real-worldpolicy improvement: reset the scene, execute a policy, verify the outcome, and refine the next iteration. To bridge this gap, we introduce ENPIRE, a harness framework forcoding agentsthat instantiates thisphysical feedback routinewith four core modules: anEnvironment module(EN) for automatic reset and verification, aPolicy Improvementmodule (PI) that launches policy refinement, aRollout module(R) to evaluate policies with one or multiple physical robots operating in parallel, and anEvolution module(E) in whichcoding agentsanalyze logs, consult literature, improve training infrastructure and algorithm code to address failure modes. This closed-loop system transforms real-world manipulation learning into a controllable optimization procedure, minimizing human effort while allowing fair ablations across training recipe and agent variants. Powered by ENPIRE, frontiercoding agentscan autonomously train a policy to achieve a 99% success rate on challenging,dexterous manipulationtasks, such as organizing a pin box, fastening a zip tie, and tool use, a process that further accelerates when we dispatch an agent team on arobot fleet. Our results suggest a practical and scalable path toward deployingcoding agentsto autonomously advancing robotics in the physical world.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2606\.19980

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2606.19980 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2606.19980 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2606.19980 in a Space README.md to link it from this page.

Collections including this paper1