HOLMES: HOLonym-MEronym based Semantic inspection for Convolutional Image Classifiers. https://t.co/mjQq1LnNwr
18,953 followers
406 followers
@amirrahnama_ I disagree — goal of XAI is to explain to unknown — how DNNs make decisions. Local attribution methods are not very useful to explain unknown behavior of the models (https://t.co/pecjdzDNO5). Global explanations allow for that (analysis of CL
29,633 followers
The Duet of Representations and How Explanations Exacerbate It. https://t.co/AzspXnMFma
29 followers
What started as a hackathon project at a summer school in 2022, is now a full-fledge article which was presented at this year's World Conference on Explainable AI and discussed during a workshop at ICML! #xaiss #xAI2023 #icml2023 #explainableAI https://