Seminar with Prof. Tomasz Hachaj on the modelling uncertainty

Date: Friday, March 5, 2021, 14:00 a.m CET

Speaker: Prof. Tomasz Hachaj, Institute of Computer Science, Pedagogical University of Krakow, Poland

Title: Modeling uncertainty in the classification process using a deep neural network: application to cosmic ray particle recognition in CMOS camera images

Abstract: The topic of the seminar will be the presentation of a model of a deep neural network allowing modeling of problems from the area of multi-label classification (classified objects may belong to several classes simultaneously). Based on the above architecture, a network has been developed that allows to classify objects, whose class labels are uncertain. The second model has been developed specifically for the recognition of cosmic ray particle images acquired in the CREDO project. The performance results of the proposed solution will also be discussed.

Language: Polish

Online access: MS Teams platform

Presentations: Slides to download and Code to download (GitHub repository)