Please Note: The content on this page is not maintained after the colloquium event is completed. As such, some links may no longer be functional.
Victor Pankratius
Computer-Aided Discovery: Scientific Insight Generation with Machine Support
Wednesday, April 19, 2017
Building 8 Auditorium - 11:00 AM
(Cookies at 10:30 AM)
The process of scientific discovery is traditionally assumed to be executed by humans. This talk highlights how increasing data volumes and human cognitive limits are challenging this traditional assumption. Relevant examples are found in geoscience, planetary science, and astronomy, disciplines that are undergoing a Big Data paradigm transformation.
This talk outlines how intelligent systems for computer-aided discovery can routinely complement human scientists. The pragmatics of model-based computer-aided discovery systems go beyond feature detection to answer more sophisticated questions, such as how empirical detections fit into hypothesized models and model variants. In particular, we facilitate the generation and exploration of connections between physics model candidates and empirical data sets to ease the scientist's work of placing large ensembles of detections into a theoretical context.
A new discovery made with this approach has recently been confirmed in volcanology by peer-review, based on data that has been collected for over a decade. Other application examples will be outlined for groundwater phenomena on a continental scale, ionospheric phenomena, solar radio bursts, exoplanet search, and planetary landing site identification on the Moon and on Mars.
Dr. Pankratius leads the Astro-& Geo-Informatics group at Massachusetts Institute of Technology, Haystack Observatory. He advances Data Science and computer-aided discovery as a principal investigator in projects supported NASA and the NSF. Throughout his computer science career he enjoyed working with students, teaching graduate and undergraduate level courses, and advising more than 30 undergraduate and graduate thesis students. At MIT, he helped create new courses on astroinformatics and geoinformatics. Contact him at pankrat@mit.edu or victorpankratius.com.
IS&T Colloquium Committee Host: Mike Little
Sign language interpreter upon request: 301-286-7040