Accelerating Discovery in Natural Science Laboratories with AI and Robotics: Perspectives and Challenges from the 2024 IEEE ICRA Workshop, Yokohama, Japan
Andrew I. Cooper, Patrick Courtney, Kourosh Darvish, Moritz Eckhoff, Hatem Fakhruldeen, Andrea Gabrielli, Animesh Garg, Sami Haddadin, Kanako Harada, Jason Hein, Maria Hübner, Dennis Knobbe, Gabriella Pizzuto, Florian Shkurti, Ruja Shrestha, Kerstin Thurow, Rafael Vescovi, Birgit Vogel-Heuser, Ádám Wolf, Naruki Yoshikawa, Yan Zeng, Zhengxue Zhou, Henning Zwirnmann
TL;DR
This viewpoint discusses accelerating discovery in natural science laboratories through AI and robotics, drawing on insights from the 2024 IEEE ICRA Workshop on Science Lab Automation. It outlines six themes: the quest and challenges of automation, digital twins and simulators, autonomy levels for flexible experiments, foundation models and generative AI for general-purpose robots, standardisation, and reproducibility, safety, sustainability, and ethics. Digital twins are proposed to reproduce experiments and provide feedback loops for optimization, with standards like SiLA2 and OPC UA enabling scalable integration; foundation models (LLMs, VLMs) support literature review, protocol generation, and multi-modal data interpretation, while RL complements classical ML for real-time decision making. The authors argue for a human-in-the-loop, modular, interoperable architecture and industry collaboration to enable broad adoption, and illustrate robustness considerations in autonomous workflows with a simple example where a 20-step process with per-step success $p=0.99$ yields end-to-end success $p^{20} \,\approx\, 0.82$.
Abstract
Science laboratory automation enables accelerated discovery in life sciences and materials. However, it requires interdisciplinary collaboration to address challenges such as robust and flexible autonomy, reproducibility, throughput, standardization, the role of human scientists, and ethics. This article highlights these issues, reflecting perspectives from leading experts in laboratory automation across different disciplines of the natural sciences.