nyuvis/explanation_explorer

The Explanation Explorer is a visual interface to explore similarly explained data items. Having a trained machine learning model it is possible to create explanations for data items by probing model inputs. The visual analytics interface groups similar explanations together and provides an interactive way of exploring the significants and quantity of those explanations in a given data set, i.e., a validation data set. You can find a live demo here which uses the example data set below. A Workflow for Visual Diagnostics of Binary Classifiers using Instance-Level Explanations; Josua Krause, Aritra Dasgupta, Jordan Swartz, Yindalon Aphinyanaphongs, Enrico Bertini; To be published at IEEE VAST 2017.