Towards a visual-based survey on explainable machine learning

The increasing use of Machine Learning in people's everyday life raised the need for solutions aimed to reveal the work done by those models when transforming an input into an output. In the field of Computer Science, techniques of Explainable Machine Learning have been developed for unveiling algorithms' inner workings at different degrees of sophistication. The current status of the research on Machine Learning Explainability is still empowering the creators of those models but is not informing the people affected by them. Being information visualization considered a good means to show these processes, it is legitimate that tools able to help designers to browse visual models used in the past are designed. The paper proposes a visual-based methodology for displaying and analyzing images in groups as a support for designers in the observation, investigation, and selection of visual models and solutions to be adopted in the area of Explainable Machine Learning.