Interpreting and Explaining Deep Models Visually

José Oramas M., Kaili Wang, Tinne Tuytelaars
2019 Belgium-Netherlands Conference on Artificial Intelligence  
Methods based on deep neural networks (DNNs) have achieved impressive results for several computer vision tasks, such as image classification, object detection, and image generation, etc. Combined with the general tendency in the community of developing methods with a focus on high quantitative performance, this has motivated the wide adoption of DNN-based methods, despite the initial skepticism due to their blackbox characteristics. Our goal is to bridge the gap between methods aiming at model
more » ... interpretation, i.e., understanding what a given trained model has actually learned, and methods aiming at model explanation, i.e., justifying the decisions made by a model.
dblp:conf/bnaic/MWT19 fatcat:bsutgtcxhffddby7auapba3m64