Using artificial intelligence and modern computing, we have an opportunity to create one of the most important innovations in the scientific method in generations, with a direct impact on critical societal challenges. We can now collapse the classic, serial process that has historically defined the scientific method into a process in which thousands or even millions of experiments are run in parallel and iterations are controlled by machine learning or optimization applied to the results. These feedback loops can permit taking on $10 worth of scientific risk for every dollar invested.
The AI-Powered Science Accelerator program promotes this new way of conducting science, including through investments in groundbreaking research projects that leverage the power of advanced computing techniques for public good.
We also drive the adoption of similar techniques in other areas of research through the establishment of a community of leaders dedicated to this idea, various scientific convenings, training programs, and the development of AI-forward science policy.
By taking a machine learning-based approach to earth modeling, a research consortium led by CalTech aims to create better climate and weather forecasting, offering the possibility of reducing economic losses from extreme weather and changing societal behavior.
Teams of researchers from UC Santa Cruz and UC San Francisco are creating millions of organoids of various parts of the brain, creating an ex vivo testbed for monitoring and testing human neural circuit behavior and unlocking the secrets of human brain development.
Researchers at MIT are using machine learning to create a tighter feedback loop in antibiotic experimentation, creating the possibility of discovering new families of antibodies that can serve as effective human therapeutics for numerous crippling diseases.
Researchers at NYU are creating a new platform to provide secure access to critical governmental data sets in areas such as employment, wages, health care, welfare assistance and others, arming policy-makers and researchers with information to change the empirical foundations of social science and improve public policy.
New Catalysts & Nanostructures
A team of scientists at the University of Washington are creating new families of protein catalysts and innovative 2- and 3-dimensional microstructures, helping to improve chemical processes, production efficiency, and reduce energy loss, as well as potentially lead to classes of pharmaceuticals.
Researchers at Princeton University and Johns Hopkins University are creating data-driven machine-learning techniques and reproducible feedback mechanisms between data collection and scientific discovery to make full use of scarce, shared telescope resources and learn more about the universe faster.