Selective Concept Models: Permitting Stakeholder Customisation at Test-Time
This work addresses the challenge of interpretability and efficiency for stakeholders in AI systems, offering a customizable approach that is incremental over existing concept-based models.
The paper tackles the problem of high cognitive load in concept-based models by proposing Selective Concept Models (SCOMs), which allow stakeholders to customize predictions using only a subset of concepts at test-time, achieving optimal accuracy with a fraction of the total concepts on multiple real-world datasets.
Concept-based models perform prediction using a set of concepts that are interpretable to stakeholders. However, such models often involve a fixed, large number of concepts, which may place a substantial cognitive load on stakeholders. We propose Selective COncept Models (SCOMs) which make predictions using only a subset of concepts and can be customised by stakeholders at test-time according to their preferences. We show that SCOMs only require a fraction of the total concepts to achieve optimal accuracy on multiple real-world datasets. Further, we collect and release a new dataset, CUB-Sel, consisting of human concept set selections for 900 bird images from the popular CUB dataset. Using CUB-Sel, we show that humans have unique individual preferences for the choice of concepts they prefer to reason about, and struggle to identify the most theoretically informative concepts. The customisation and concept selection provided by SCOM improves the efficiency of interpretation and intervention for stakeholders.