Accelerating Discovery in Natural Science Laboratories with AI and Robotics

About 

Large-scale, high-throughput experiments, and big data are becoming fundamental for scientific discoveries. New discoveries are time-consuming and costly; hence, laboratory automation based on robotics and AI aims to shorten and reduce these factors. Most scientists (Baker, 2016) believe that the lack of reproducibility and scalability of experiments is the main reason for this situation, especially in the fields of life sciences, material, and drug discovery. For years, automating scientific experiments has been considered the holy grail for addressing this fundamental reproducibility problem. Most of the existing solutions are limited to rigid and complex devices that can only solve specific experimental tasks that are barely adjustable to experimental protocol changes. With further developments in robotics and intelligent software, new opportunities are opening up to address this in a more task-flexible but human-centric way. However, it is not yet clear how these developments can be applied to accelerate laboratory processes. Specific challenges e.g., manipulation and perception of transparent glassware, system integration with non-digital devices, challenges with deploying mobile manipulators safely in human labs, as well as standardization, have not yet been fully overcome. This workshop will provide an opportunity to create new synergies toward addressing these challenges. 

The main objective of the workshop is to build synergies between these communities (robotics and natural sciences), where the main aim is to focus on current open challenges in using robotics and AI methods in semi-structured laboratory environments.