BioAgent robot learns and adapts to complex lab tasks autonomously

BioAgent: Self-Directed Skill Acquisition and Capability Growth for Long-Horizon Laboratory Protocols

Robotics

Summary

Robots often struggle to follow complicated lab instructions that are new to them. The authors present BioAgent, a robot that can figure out what skills it lacks, learn new ones by asking humans, and check if the steps in an experiment are still okay as it works. BioAgent tests what it has learned to make sure each step’s result fits and can correct itself if something goes wrong. When tested on two difficult lab tasks involving many steps, it successfully completed the tasks more than half the time, showing it can grow its abilities over time.

What this means in practice

  • For laboratory automation teams: Improve laboratory robots’ ability to autonomously learn new experimental steps and recover from errors during complex protocols.
  • For industrial automation developers: Develop robots that can adapt to new multi-step manufacturing processes by acquiring missing skills and validating sequential task outcomes.

Authors

Gaojing Zhang, Changwen Chen, Zichen Luo, Jialiang Li, Yuhan Wang, Yifei Xie, Wenzhao Lian

Abstract

Executing an unfamiliar laboratory protocol requires a robot to identify missing capabilities and determine whether each operation remains valid as the experiment progresses. This letter presents BioAgent, an embodied agent that addresses these coupled challenges of capability sufficiency and experimental validity. It matches human-reviewed requirements to validated skills and task packages, selects tool-based implementations for missing behaviors, and requests human teaching when needed during learning. Learned skills undergo sequential task-package validation to check outcomes and compatibility between successive operations. During execution, precondition, outcome, and experimental-state checks govern progression, autonomous recovery, or termination. Execution evidence guides tool selection, while accepted experience supports offline policy updates and task packaging. On the second task, BioAgent reuses 17 skill types unchanged, revises two existing implementations, and acquires five new types. On two dual-arm laboratory tasks with 20 and 31 ordered skill invocations, BioAgent achieves complete-task success rates of 66\% and 58\% with autonomous recovery over 50 runs per task. Across 25 atomic skills evaluated over 50 autonomous trials each with recovery disabled, success rates range from 84\% to 100\%, with 17 skills achieving 50/50.