Autonomously Acquiring Robot Manipulation Skills with Language-Driven Quality-Diversity
2026-08-31 • Robotics
Robotics
AI summaryⓘ
The authors address the challenge of teaching robots to perform a variety of different movements without needing experts to design specific rules for success or diversity. They developed a method that uses natural language task descriptions and large language models (LLMs) to automatically explore and generate many diverse robot motions. Their approach treats the search for useful behaviors as a simpler functional design problem and adapts existing quality-diversity algorithms to handle these designs. Experiments showed their method creates more diverse motion sets for robotic tasks compared to traditional methods that rely on expert-written rules.
Quality-Diversity algorithmsMotion primitivesRobot learningLarge Language Models (LLMs)Fitness functionBehavior descriptorsPolicy explorationMAP-Elites algorithmReward shapingAutonomous exploration
Authors
Émiland Garrabé, Mahdi Khoramshahi, Stéphane Doncieux
Abstract
Quality-diversity (QD) algorithms have been gaining traction in robot learning, where diverse motion primitive libraries allow robots to adapt zero-shot to constraints at deployment time. However, such methods typically require expert designers to write the success condition, fitness and diversity metrics, and this strongly limits the robot's autonomy. On the other hand, existing LLM-based reward-shaping techniques allow robots to learn autonomously but only output single high-performing solutions, limiting the robot's adaptability. In this paper, we propose an approach designed to output diverse motion primitive archives by autonomously leveraging quality-diversity algorithms, only requiring a free-form description of the task in common language. To address the difficulty of designing relevant fitness and diversity metrics, we propose an autonomous exploration mechanism able to reliably output sets of functionals covering the fitness and behavior descriptor (BD) space. First, we pose policy exploration as a functional design problem, where the functional spaces are lower-dimensional than the full BD and fitness spaces, and propose an LLM-based exploration scheme to sample from these low-dimensional spaces without any task-specific prompts, fine-tuning or expert intervention. We adapt a multi-BD variant of the MAP-Elites success (MES) algorithm, designed to leverage the heterogeneous BD samples. Finally, through experiments based on the genesis simulator, we show that our method effectively generates archives of diverse motion primitives, outperforming classical QD algorithms with inferred and hand-written parametrizations on a set of $4$ robotic manipulation tasks.