Digging Deeper: Learning Multi-Level Concept Hierarchies
This addresses the problem of limited interpretability in concept-based models for AI researchers and practitioners by enabling deeper concept hierarchies without exhaustive annotations, representing an incremental advancement over prior shallow-hierarchy methods.
The paper tackles the limitation of shallow concept hierarchies in interpretable AI by introducing Multi-Level Concept Splitting (MLCS) to discover multi-level concept hierarchies from only top-level supervision and Deep-HiCEMs to represent these hierarchies, showing that MLCS discovers human-interpretable concepts absent during training and Deep-HiCEMs maintain high accuracy while enabling test-time interventions that improve task performance.
Although concept-based models promise interpretability by explaining predictions with human-understandable concepts, they typically rely on exhaustive annotations and treat concepts as flat and independent. To circumvent this, recent work has introduced Hierarchical Concept Embedding Models (HiCEMs) to explicitly model concept relationships, and Concept Splitting to discover sub-concepts using only coarse annotations. However, both HiCEMs and Concept Splitting are restricted to shallow hierarchies. We overcome this limitation with Multi-Level Concept Splitting (MLCS), which discovers multi-level concept hierarchies from only top-level supervision, and Deep-HiCEMs, an architecture that represents these discovered hierarchies and enables interventions at multiple levels of abstraction. Experiments across multiple datasets show that MLCS discovers human-interpretable concepts absent during training and that Deep-HiCEMs maintain high accuracy while supporting test-time concept interventions that can improve task performance.