Coffee Hour with Charles Twardy: Data Science for Search and Rescue

Share
Time: 
Friday, February 24, 2017 - 3:30pm
Place: 
Refreshments are offered in 319 Walker Building at 3:30 p.m. The lecture begins in 112 Walker Building at 4:00 p.m.

Euclidean and Watershed distance model

The image shows the combined Euclidean and Wastershed distance model for hiker in a flat and dry environment. Dark shares represent higher probability density.

This week's Coffee Hour lecture is co-sponsored by the Penn State Institute for CyberScience.

About the talk

Lost-person search is a mystery with a deadline. After 24 hours lost in the wilderness, your survivability drops by 20%. Searches happen one at a time, but nationally they consume thousands of hours and millions of dollars per year. Most of the expense is borne by ~5% the searches. These massive, extended searches would benefit from proper application of Bayesian search theory, developed in WW2 and used successfully by the Navy and Coast Guard for the past 70 years. I will discuss the unique challenges in wilderness search, and the progress made since 2000 due to the worldwide collection of lost-person data. I’ll discuss spatial probability maps for lost-person behavior, survival curves lost persons, and empirical and theoretical detection profiles, and the prospects for Bayesian search management.

About the speaker

Charles TwardyCharles R. Twardy led the George Mason team in the IARPA ACE and ForeST forecasting programs from 2011-2015, beating the baseline by about 35 percent and creating a new kind of large-scale combinatorial prediction market. From 2015 to 2016 he worked as Senior Data Scientist for NTVI Federal in support of the Defense Suicide Prevention Office. In late 2016 he joined Sotera where he works on the DARPA MEMEX and IARPA CREATE programs. Twardy’s research focuses on evidence and inference with a special interest in causal models, Bayesian networks, and Bayesian search theory, especially the analysis and prediction of lost person behavior. He also has worked on argument mapping, information-theoretic trajectory clustering an sensor-selection, and Bayesian models for counter-IED work, source credibility, image recognition, environmental decision-making, and epidemiology. He received a Dual Ph.D. in Cognitive Science and History & Philosophy of Science from Indiana University, and a B.A. from the Interdisciplinary Majors program at the University of Virginia.

Suggested readings

Contact us

Penn State encourages qualified persons with disabilities to participate in its programs and activities. If you anticipate needing any type of accommodation or have questions about the physical access provided, please contact Angela Rogers in advance of your participation or visit.

Angela Rogers  office: 814-865-2493 email: geography@psu.edu

Coffee Hour