Time: 15:00 on 24/10/2018
Place: room 1002, building B1 (HUST)
Presenter: Nguyen Vu Linh, UMR CNRS 7253 Heudiasyc, Sorbonne universités, Université de technologie de Compiègne, Compiègne, France
Title: Epistemic and Aleatoric Uncertainties in Active Learning and Classification
Abstract:
The distinction between the epistemic and aleatoric uncertainty involved in the predictions of new instances is well-accepted in the literature on uncertainty and has been considered in only few recently
machine learning works. Roughly speaking, the aleatoric uncertainty refers to the notion of randomness, that is, the variability in the outcome of an experiment which is due to inherently random effects. As
opposed to this, the epistemic uncertainty refers to the uncertainty caused by a lack of knowledge. Yet, the distinction epistemic/aleatoric uncertainty has been well-studied in the literature. To facilitate subsequent machine learning applications, we have developed practical procedures to estimate these degrees for popular classifiers. In particular, we have explored the use of this distinction in the contexts of active learning and cautious inferences
Bio: Vu-Linh Nguyen is a postdoctoral researcher at the Paderborn University, Germany. He received his B.S. degree in Mathematics from the VNU University of Science, Vietnam, M.S. and Ph.D. degrees in Computer Science from the Japan Advanced Institute of Science and Technology, Japan, and the University of Technology of Compiègne, France. He has been working in machine learning. His current interests are uncertainty modeling, where available data or knowledge suffers from important imperfections: partially specified data, randomness, lack of information, with applications in classification and active learning.