In our research group we aim to reduce the time and money required to discover a new functional material or optimize a known one by a factor of ten, namely from ran estimated ten million dollars and ten years of development to one million dollars and one year. We suggest that the solution for this challenge is the development of self-driving laboratories.
Self-driving laboratories combine artificial intelligence with automated robotic platforms for the autonomous discovery of new materials. The Aspuru-Guzik group sees in these laboratories the potential to increase the rate of experimentation and scientific discovery, which will eventually change the way we do science. The concept of interconnected autonomous robotic platforms with specific tools to tune materials properties allows accessing the realm of novel fabrication strategies.
Creating a fully autonomous self-driving lab is a multidisciplinary task that combines a diverse set of research fields. Machine learning and modeling methods are used for predicting materials properties and suggesting new experiments. While robotics, computer vision, and automated characterization methods are used to perform the experiments and analyzing the results. A key element for designing autonomous laboratories consists of closing the loop between experiments and computational modeling by bridging these independent technologies into a single platform.
The Self-driving lab subgroup explores a diverse range of fields, including Artificial intelligence and optimization methods for controlling and designing experiments. Robotics systems for performing these experiments, and automatic characterization methods for analyzing the results.
A novel research direction is the use of computer vision for the creation of visually-aware robotic systems that can carry out chemical and materials science experiments.