Last December, the Acceleration Consortium (AC) began exploring the use of Merck KGaA, Darmstadt, Germany’s AI-driven experimentation planner, Bayesian Back End (BayBE). BayBE is being benchmarked by the AC’s polymer and formulation labs.
“Since our collaboration announcement, the AC began exploring this tool,” said Alán Aspuru-Guzik, director of the Acceleration Consortium. “Bayesian optimization enables researchers to speed up and reduce the cost of the discovery of new materials and chemical processes by making the experimental process more efficient and informed. We have been putting BayBE to the test and comparing it to other solutions, including our own codes, I am excited to see how industry and academia can work together to accelerate finding solutions to challenges in chemistry and materials science, which is what the AC is all about.”
One of these tests was a hackathon in late March 2024. The hackathon, sponsored by the Acceleration Consortium and Merck KGaA, brought 120 participants from academia, industry, and government together to give them the chance to test BayBE. Participants selected or developed Bayesian optimization algorithms and applied them to benchmarking tasks, designed new benchmark tasks, created instructional tutorials and described real-world Bayesian optimization applications.
“Bayesian optimization is like having a smart assistant that tells you the most promising experiments to run next,” said Dr Sterling Baird, Director of Training and Programs at the Acceleration Consortium. “It helps in research where there is uncertainty, or when outcomes aren’t straightforward or when there are a lot of variables. For this hackathon, we set challenges for the participants that had all these aspects in them to test their skills and encourage creative outcomes.”
The AC has also hosted collaboration-driven events such as workshops and seminars to introduce open source software such as BayBE to the global AC network. This strategic alliance underscores the significance of these tools and their role in advancing the mission of the AC in the realm of self-driving labs.
“BayBE is a game changer for chemical applications in research and exploration," said Philipp Harbach, Global Head of Group Digital Innovation at Merck KGaA, Darmstadt, Germany. "Not only does it increase efficiency, but it consistently delivers superior results and fosters breakthrough solutions. Its impact spans multiple chemical domains, from small drug-like molecules and organic chemistry to formulations and polymers, positioning BayBE as a standard for automation. By collaborating with the Acceleration Consortium, we are paving the way for seamless integration of BayBE enabling fully automated experimentation cycles that drive innovation across the board.”
“As an industry partner of the Acceleration Consortium since its founding, it has been exciting to see creative models of collaboration towards progressing the field of AI and automation for materials discovery together," said Jacob Woodruff, Head of Technology Scouting and Partnerships with Merck KGaA, Darmstadt, Germany.
