Abstract:
Today AI tools are largely used for information analysis and decision making. For critical applications such as healthcare, security… the users of AI tools need to understand why the system has taken such a decision.
Image classification is one of the most frequent problems AI tools such as Deep Neural Networks are designed for. The explanation methods aim to highlight elements in the input accordingly to their contribution to the final decision.
In case of image classification tasks we are speaking of pixel importance or explanation maps. The methods of their building can be divided into two categories: the so-called black-box methods and white -box methods.
In the talk we will present the taxonomy of the main explainers nowadays. We will go into details of some methods developed by us and also speak about evaluation of explainers, which is an open research question per se.
About the speaker:
Jenny Benois-Pineau is a full professor of Computer Science at the University Bordeaux and chair of Video Analysis and Indexing research group in Image and Sound Department of LABRI Université Bordeaux. She is also chair of International relations at School of Science and Technology of University of Bordeaux comprising 9.500 students.
Professor Benois-Pineau’s topics of interest include image and video analysis, artificial intelligence in multimedia and healthcare. She is the author and co-author of more than 200 papers in international journals, conference proceedings, books and book chapters.
She is associate editor of Signal Processing: Image Communication (Elsevier) and Multimedia Tools and applications (Springer) journals and senior associate editor of Journal of Electronic Imaging (SPIE) journal.
She has served in numerous program committees in international conferences and organised workshops such as ACM MM, CIVR, CBMI, EI, MMM, IEEE ICIP, IEEE/IAPR ICPR and gave invited lectures in the universities of USA, Europe, Israel and Mexico.
She is a member of IEEE TC on IVMSP. She was decorated by Knight of Academic Palms grade.