LabEvolver: Training-Free Experience Evolution for Safe and Grounded Wet-Lab Agents

2026/07/30 by Jingya Wang, Yuyang Gao, Liuzhenghao Lv +2
Computer Science · #cs.RO #cs.AI

paper · pdf

arxiv created 2026/07/30 · arxiv updated 2026/07/31

Abstract

We introduce LabEvolver, a training-free framework that equips safe and grounded wet-lab agents with episodic memory from execution experience. LabEvolver couples a state-grounded inner trial loop for adaptive perception, online planning, and safety validation with an outer evolution loop that distills completed trajectories into reusable skill, strategy, and safety experience. On robotic solution-preparation tasks, LabEvolver demonstrates real-world feasibility, reducing pH-regulation completion time and safety-gate intercepts by 48.2% and 60.0%, respectively. On ALFWorld, it further improves cumulative success rate within 20 steps from 76.2% with ReAct to 91.4% over 500 continual tasks, showing generality beyond wet-lab settings. These results support learn-by-doing experience evolution as a feasible path toward closed-loop automated scientific discovery. The project page is available at https://github.com/AndyGao6186/LabEvolver.

Citations

• Memp: Exploring Agent Procedural Memory
• G-Memory: Tracing Hierarchical Memory for Multi-Agent Systems
• DROID: A Large-Scale In-The-Wild Robot Manipulation Dataset
• Empowering Large Language Model Agents through Action Learning
• ORGANA: A Robotic Assistant for Automated Chemistry Experimentation and Characterization
• Autonomous chemical research with large language models
• Open X-Embodiment: Robotic Learning Datasets and RT-X Models
• RoboAgent: Generalization and Efficiency in Robot Manipulation via Semantic Augmentations and Action Chunking
• ExpeL: LLM Agents Are Experiential Learners
• RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control
• Voyager: An Open-Ended Embodied Agent with Large Language Models
• ChemCrow: Augmenting large-language models with chemistry tools
• Self-Refine: Iterative Refinement with Self-Feedback
• Reflexion: Language Agents with Verbal Reinforcement Learning
• PaLM-E: An Embodied Multimodal Language Model
• RT-1: Robotics Transformer for Real-World Control at Scale
• ReAct: Synergizing Reasoning and Acting in Language Models
• ProgPrompt: Generating Situated Robot Task Plans using Large Language Models
• Code as Policies: Language Model Programs for Embodied Control
• Do As I Can, Not As I Say: Grounding Language in Robotic Affordances
• ALFWorld: Aligning Text and Embodied Environments for Interactive Learning
• Self-driving laboratory for accelerated discovery of thin-film materials
• Organic synthesis in a modular robotic system driven by a chemical programming language

Related

• Pipette: An Embodied Simulation Platform, Benchmark, and Data-Efficient Augmentation Framework for Wet-Lab Robotics
• FLEX: Continuous Agent Evolution via Forward Learning from Experience
• FlowEvo: Self-Evolving Agents through the Co-Evolution of Workflows and Executable Skills
• LabRobFail: A Benchmark for Robotic Failure Analysis in Chemical Self-driving Laboratory
• LabUtopia: High-Fidelity Simulation and Hierarchical Benchmark for Scientific Embodied Agents