aiqbench analysis
ASPIRE introduces a significant advancement in robotic automation by utilizing large language models to autonomously write, debug, and refine control programs through an iterative learning loop. This system effectively addresses the complexity of multimodal perception and physical dynamics, enabling zero-shot generalization on unseen long-horizon tasks with performance that scales directly with frontier model capabilities. The approach represents a pivotal shift from manual programming to agentic skill discovery, substantially reducing the effort required for real-world robot deployment.
ASPIRE: Agentic /Skills Discovery for Robotics
arXiv:2607.00272v2 Announce Type: replace Abstract: Traditional robot programming is notoriously challenging: it requires orchestrating multimodal perception, managing physical contact dynamics, and handling diverse environment configurations and execution failures. We introduce ASPIRE (Agentic Skill Programming through Iterative Robot Exploration), a continual learning system that autonomously writes and refines robot control programs in a code-as-policy paradigm while compounding experience into a reusable skill library. ASPIRE operates in an open-ended learning loop, consisting of three components: (1) a closed-loop robot execution engine that exposes fine-grained multimodal traces, enabling the agent to autonomously diagnose failures, synthesize repairs, and validate outcomes; (2) a continually expanding skill library that distills validated fixes into reusable, transferable robotic knowledge; and (3) an evolutionary search procedure that generates diverse programs, systematically debugging them to explore beyond single-trajectory refinement. ASPIRE surpasses prior methods by up to 77% on manipulation tasks under perturbation, 72% on Robosuite's bimanual handover task, and up to 32% on long-horizon household tasks. The accumulated skill library further enables zero-shot generalization: on representative unseen long-horizon tasks, ASPIRE achieves 31% success, outperforming the 4% success rate of prior methods. Performance scales with the capabilities of frontier models. With Claude Opus 5.5, ASPIRE nearly saturates these benchmarks, achieving 99% average success across all 80 LIBERO-Pro tasks and 100% success on representative Robosuite and BEHAVIOR-1K tasks. Beyond simulation, skills discovered in simulation reduce real-world robot programming effort across embodiments and robot APIs, while skills acquired through online interaction with physical robots enable ASPIRE to solve a Rubik's Cube with progressively improving efficiency.
Abstract:Traditional robot programming is notoriously challenging: it requires orchestrating multimodal perception, managing physical contact dynamics, and handling diverse environment configurations and execution failures. We introduce ASPIRE (Agentic Skill Programming through Iterative Robot Exploration), a continual learning system that autonomously writes and refines robot control programs in a code-as-policy paradigm while compounding experience into a reusable skill library. ASPIRE operates in an open-ended learning loop, consisting of three components: (1) a closed-loop robot execution engine that exposes fine-grained multimodal traces, enabling the agent to autonomously diagnose failures, synthesize repairs, and validate outcomes; (2) a continually expanding skill library that distills validated fixes into reusable, transferable robotic knowledge; and (3) an evolutionary search procedure that generates diverse programs, systematically debugging them to explore beyond single-trajectory refinement. ASPIRE surpasses prior methods by up to 77% on manipulation tasks under perturbation, 72% on Robosuite's bimanual handover task, and up to 32% on long-horizon household tasks. The accumulated skill library further enables zero-shot generalization: on representative unseen long-horizon tasks, ASPIRE achieves 31% success, outperforming the 4% success rate of prior methods. Performance scales with the capabilities of frontier models. With Claude Opus 5.5, ASPIRE nearly saturates these benchmarks, achieving 99% average success across all 80 LIBERO-Pro tasks and 100% success on representative Robosuite and BEHAVIOR-1K tasks. Beyond simulation, skills discovered in simulation reduce real-world robot programming effort across embodiments and robot APIs, while skills acquired through online interaction with physical robots enable ASPIRE to solve a Rubik's Cube with progressively improving efficiency.
41 pages, 19 figures, 26 tables. Project page: this https URL Code: this https URL
Robotics (cs.RO); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
arXiv:2607.00272 [cs.RO]
(or arXiv:2607.00272v2 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2607.00272 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 359 words
Related coverage
Motion Primitive Discovery in a Humanoid Robot via Self-Organising Maps for Phase Recognition
arXiv · Jul 22, 2026
Betting on AI and Robots to Automate Superconductor Discovery
spectrum.ieee.org · Sep 8, 2026
Multi-Dimensional Featurizers Yield Fine-Grained Steering and Data Shaping in Robot Foundation Models
arXiv · Oct 16, 2026
What Stops Recursive Self-Improvement in Robotics? Lessons from 123 Rounds of Agentic Skill Discovery
arXiv · Sep 30, 2026