AI4science
Across science—from astrophysics to molecular biology to economics—researchers are overwhelmed by the sheer amount of data they are collecting. But, this problem is better viewed as an opportunity since, with the right computing resources and algorithmic tools, scientists might unlock new insights from the swathes of data to carry their field forward.
AI4science (or artificial intelligence for science) is an initiative at Caltech led by Anima Anandkumar and Yisong Yue that aims to bring together AI researchers with experts from other disciplines to push modern AI tools into every area of science and engineering. Launched in the summer of 2018, the initiative organizes talks, courses, and tutorials aimed at training researchers from across the scientific spectrum in the theory and practice of machine learning algorithms. Weekly AI4science office hours also allow researchers to ask computer scientists for help—hopefully stimulating new interdisciplinary research at Caltech. Additionally, seed grants for researchers applying AI to new applications in the sciences are allocated yearly through the Carver Mead New Adventures program.
Here are a few examples of Caltech professors transferring artificial intelligence and machine learning research to other disciplines.
To find out more about this new initiative, visit the AI4science website.
AI4neuro
Joel Burdick, Richard L. and Dorothy M. Hayman Professor of Mechanical Engineering and Bioengineering, is applying machine learning algorithms (some designed by Caltech's Professor Yisong Yue) to help patients with spinal cord injuries to walk again. The so-called "neural prosthesis" is a device which plugs into a human's nerve endings, reads the electrical signals and uses them to control a mobility-assisting device.
AI4climate
Andrew Stuart, Bren Professor of Computing and Mathematical Sciences, is applying machine learning algorithms to build more fine-grained climate models. Working with Professor Tapio Schneider in Caltech's Climate Dynamics Group, Stuart hopes to build better predictive models of the Earth's changing climate.
AI4physics
Maria Spiropulu, Professor of Physics, is applying machine learning methods to try to make sense of the floods of data coming in from the Large Hadron Collider at CERN. Currently most of the roughly one petabyte of data collected every second at CERN must be thrown away, so even deciding which data to keep is an important problem in the hunt for signals in the dark.