Discovering and Validating AI Errors With Crowdsourced Failure Reports
This addresses the issue of scaling failure detection for AI practitioners, though it is incremental in applying crowdsourcing and visualization to an existing bottleneck.
The paper tackles the problem of discovering systematic AI failures by introducing crowdsourced failure reports and a visual analytics system called Deblinder, which helps developers detect and validate errors, and shows that collecting additional data from identified groups can improve model performance.
AI systems can fail to learn important behaviors, leading to real-world issues like safety concerns and biases. Discovering these systematic failures often requires significant developer attention, from hypothesizing potential edge cases to collecting evidence and validating patterns. To scale and streamline this process, we introduce crowdsourced failure reports, end-user descriptions of how or why a model failed, and show how developers can use them to detect AI errors. We also design and implement Deblinder, a visual analytics system for synthesizing failure reports that developers can use to discover and validate systematic failures. In semi-structured interviews and think-aloud studies with 10 AI practitioners, we explore the affordances of the Deblinder system and the applicability of failure reports in real-world settings. Lastly, we show how collecting additional data from the groups identified by developers can improve model performance.