Scalable, Multi-output Gaussian Processes (GPs) & Applications

 Scalable, Multi-output Gaussian Processes (GPs) & Applications
12:30-2pm  hosted by Dr. Hyungusk Tak – 538 Davey Lab

We will have short talks by researchers who will provide an overview of potential use cases for scalable and multi-output GPs in various fields, as well as a generally applicable subsampling method for scalable inference. We will then open the floor for discussion about their pros and cons, general applicability, and exchangeability across fields, particularly with respect to problem size and the scaling of computational and memory costs. We are especially interested in identifying researchers with relevant expertise across multiple disciplines and exploring potential opportunities for collaborative proposals. This workshop is jointly sponsored by the Center for Astrostatistics & Astroinformatics and ICDS.

12:30pm –  Hyungsuk Tak (Statistics, Astronomy, ICDS, CAStAI): Scalable, Multi-output GPs in Astronomy.
12:45pm –  Samuel Baugh (Statistics): Scalable, Multi-output GPs in Spatio-temporal Statistics.
1:00pm –  Justin Silverman (Statistics, Medicine, ICDS): Scalable, Multi-output GPs in Bio-medical Sciences.
1:15pm –  Le Bao (Statistics): Subsampling as a General Tool for Scalable Inference
1:30 –  Q&A with Speakers