🟠Topic：

Accounting for Unknown Unknowns and Black Swans in Data Science

在資料科學中考量無法預測的未知與黑天鵝事件

🗣️Speaker：

Robert B. Gilbert, PhD, PE, D.GE, NAE

Chair, Department of Civil, Architectural and Environmental Engineering, The University of Texas at Austin

Member, U.S. National Academy of Engineering

Member, U.S. Army Corps of Engineers Post-Hurricane Katrina Evaluation and Recovery Committee

📌Abstract：

Probability theory, and specifically Bayes’ Theorem, is the foundation for artificial intelligence and information processing. A persistent challenge in applying Bayes’ Theorem is establishing the starting point before information is incorporated, which is a comprehensive set of all possible events and their respective probabilities. How do we include all possibilities and assess probabilities a priori without any information? How do we account for an unknown that is outside our range of experience? How do we include the possibility of black swans when we have only seen white swans?

This talk will describe our research to establish a non-informative starting point in applying Bayes’ Theorem in a rational and defensible way, Decision Entropy Theory. The premise of this theory is that non-informative probabilities will not inform a decision. The theory establishes a mathematical framework that describes the entropy (or uncertainty) of a decision, and it finds non-informative prior probabilities by maximizing the entropy of the decision.

This talk will provide practical examples, including the deadly 2014 landslide in Washington state, to illustrate the challenge and importance of establishing non-informative prior probabilities and to illustrate how Decision Entropy Theory attempts to address this challenge.

📅Information :

Date & Time: Wednesday, July 2nd, 10:00 AM – 12:00 PM

Venue: International College Salon, 7th Floor, Zonghe Lecture Building, National Taiwan University (NTU)

Language: English

Registration Link: https://forms.gle/v8wwfbkQ654xwoWu7

Contact Emails:

Ms. Jing-Qi Loh – jingqiloh1998@ntu.edu.tw

Professor Kuo-Hsin Yang – khyang@ntu.edu.tw