Recently, Vasileios Mezaris, Research Director at the Centre for Research and Technology Hellas, Information Technologies Institute, presented some of the technological aspects of the work CETH-ITI is conducting within the scope of CRiTERIA project in a talk titled “Explaining the Decisions of Image/Video Classifiers”.
The presentation was part of the agenda of the 1st Nice Workshop on Interpretability, November 17-18, 2022, Université Côte d’Azur, Nice, France. The two-day workshop’s aim was to create links between researchers working on the interpretability of machine learning models, in a broad sense.
You can download the slides from Vasileios Mezaris’ presentation in our publications portal.
About the "Explaining the Decisions of Image/Video Classifiers" Session
The talk started by discussing the main classes of explainability approaches for image and video classifiers. Then, the focus was shifted to two distinct problems: learning how to derive explanations for the decisions of a legacy (trained) image classifier, and designing a classifier for video event recognition that can also deliver explanations for its decisions. Technical details of the proposed solutions to these two problems will be presented. Besides quantitative results concerning the goodness of the derived explanations, qualitative examples will also be discussed in order to provide insights on the reasons behind classification errors, including possible dataset biases affecting the trained classifiers.
Download the Latest Publications
Visit our publications portal to see all of the recently published scientific papers and presentations about the work we do in CRiTERIA.