Visual, textual or hybrid: the effect of user expertise on different explanations
summary
Two between-subject studies (each with 10 experts and 10 non-experts) compare a textual explanation, a visual explanation using partial dependence plots, and a combined hybrid explanations design. Non-experts prefer visual explanations but answer significantly fewer insight tasks correctly and show confirmation bias and reduced exploration; adding text raises correct interpretation and exploration for non-experts and gives experts reassurance, so the hybrid materially reduces the expertise gap.
AI-written summary · about
related talks
22mPersonalized explanations for hybrid recommender systems
12mExplaining recommendations in an interactive hybrid social recommender
26mThe effect of explanations and algorithmic accuracy on visual recommender systems of artistic images
18mDecision making strategies differ in the presence of collaborative explanations: two conjoint ...
16mUser Interaction Models for Disambiguation in Programming by Example
41mIf We Want AI to be Interpretable, We Need to Measure Interpretability
37mInteractive Explainable AI
19mProgressive disclosure: empirically motivated approaches to designing effective transparency