Dive into BayBE – a new AI-driven opensource experimentation planner launched by Merck KGaA, Darmstadt, Germany with Acceleration Consortium
Subscribe to our mailing list, visit our website, or follow us on Twitter and LinkedIn.
Merck Group’s Digital Chemistry team and the Acceleration Consortium at University of Toronto proudly announce the launch of a new AI-driven experimental planner, Bayesian Back End (BayBE), powered by artificial intelligence and scientific acumen.
Available open-source on Github with a special emphasis on chemistry and materials science, the goal is to encourage widespread adoption and foster an inclusive scientific community, inviting researchers and innovators from diverse backgrounds to contribute and shape the future of AI-driven experimentation.
Collaboration is a guiding principle of the Acceleration Consortium. “We believe collective expertise brings transformative power to the forefront of innovation. Together, with our partners, we’re merging industry insights with academic research to fast-track discoveries. The most rewarding part of all is opening solutions to the greater scientific community. Join us in this collaborative journey—explore this opensource planner that will help inspire the next wave of scientific breakthroughs.” - Padraic Foley, Director, Strategy & Partnerships at Acceleration Consortium
BayBE serves as a general-purpose toolbox for iterative experimentation, utilizing Bayesian optimization to provide systematic recommendations that will expedite improved experiments, reduce costs, and accelerate scientific discoveries.
The release of BayBE as open-source software signifies a significant leap toward collaborative innovation, where the fusion of expertise from Merck KGaA, Darmstadt, Germany, and the Acceleration Consortium promise to reshape the landscape of self-driving labs.
Read the full story on our website.
Subscribe to our mailing list, visit our website, or follow us on Twitter and LinkedIn.
If you liked this post from A Materially Different Newsletter, why not subscribe?