AnchorViz: Facilitating Classifier Error Discovery through Interactive Semantic Data Exploration

In supervised interactive machine learning, human knowledge about the target concept can be a powerful reference to build a concept classifier that is robust to unseen items in the real world. The main challenge lies in finding unlabeled items that can either help discover or refine subconcepts for which the current classifier has no corresponding features (i.e., it has feature blindness). Yet it is unrealistic to ask humans to come up with an exhaustive list of items, especially for rare subconcepts that are hard to recall. This paper presents AnchorViz, an interactive visualization that facilitates error discovery through semantic data exploration. By creating example-based anchors, users create a topology to spread data based on their similarity to the anchors and examine the inconsistencies between data points that are semantically related. The results from our user study show that AnchorViz helps users discover more prediction errors than stratified random and uncertainty sampling methods.