When
Where
Dr. Erica Briscoe from Georgia Tech Research Institute is giving a research talk Guided Tools for Scientific Exploration" as part of the School of Information's Colloquium Series.
Abstract: The increasing sophistication of NLP and computational reasoning mechanisms are positioning researchers to be able to automate and improve many of the traditionally laborious tasks that are critical parts of the scientific research process. In this talk we discuss our design, implementation and evaluation of two platforms that aim to better scientific research and discovery. The first, TechSight, is designed to automate the discovery and characterization of progress across science and technology research, producing analytics over varied datasets, including journal publications, patents, venture capital investments, and government program funding. We discuss our testing of various algorithmic approaches (such as densification-based metrics for bibliographic networks) for quantifying scientific evolution, especially as focused on the discovery of scientific surprise and disruptive technology. The second tool is Archean – designed to aid and augment literature review and experimental study design for social scientists. We discuss how users, both novice and experienced, make use of these analysis tools in order to discover and document research in science and technology and present lessons learned from the design of scientific research-focused analytics platforms.
Biography: Dr. Erica Briscoe is the interim Chief Technology Officer and a lab Chief Scientist at the Georgia Tech Research Institute, the applied research arm of Georgia Tech. As CTO, she oversees a large portfolio of internal research and development (IRAD) programs spanning from robotics to hypersonics. In addition to her role as CTO, she works with a diverse group of researchers to conduct basic research and develop operational behavioral and data science/analytics applications in various areas. Recent and current projects involve research and software development centered on the creation and application of analytics to open source and Multi-INT, which marries computational work across many fields (including natural language processing, graph analytics, and behavior modeling) to produce actionable intelligence in various domains, such as scientific discovery and information operations. Other projects focus on a number of problem spaces, including: computational social science, technology emergence and prediction, social network analysis, insider threat detection, terrorism and radicalization, human and machine perception, the characterization of statecraft, and human concept formation. She has been funded by DARPA, NIJ, CDC, ASD R&E, DTRA, and NSF. Dr. Briscoe received a BS degree in Industrial Engineering from Georgia Tech, an MS degree in Information Science from Drexel University, and an MS and PhD from Rutgers University in Cognitive Psychology.