"Progress in AI for science and society"

Abstract:

Machine learning and reasoning systems have undergone a revolution in capabilities over the past decade, driven by increasing use and availability of compute and data. One key area of application for these new technologies is in advancing the forefront of science. In the first half of my talk, I will discuss research efforts over the past 8 years to integrate these tools into the calibration and running of complex physics experiments using semiconductor quantum dots, including most recently completely automated tune-up from cool down to quantum gates. At the same time, these tools come with increasing risk for their misuse and second-order unintended consequences of their action. To better enable a path that benefits society, I will discuss some efforts to create a metrological science around modern AI and to integrate a community of practice through the creation of the U.S. AI Safety Institute and its Consortium.

Bio:

Jake Taylor is a NIST Fellow, a Fellow of the Joint Quantum Institute (JQI), and co-founder and Fellow of the Joint Center for Quantum Information and Computer Science (QuICS). He served from 2017-2020 as the first Assistant Director for Quantum Information Science at the White House Office of Science and Technology Policy (OSTP), and founded and led the National Quantum Coordination Office (quantum.gov) from 2019-2020.