Concept Embedding Models: Beyond the Accuracy-Explainability Trade-Off
This work addresses the need for trustworthy AI systems that support human interactions by improving concept bottleneck models, offering a novel method that is incremental in advancing the field of interpretable machine learning.
The paper tackles the problem of concept bottleneck models struggling to balance high task accuracy, robust explanations, and effective interventions, especially with scarce concept supervision, by proposing Concept Embedding Models that learn interpretable high-dimensional concept representations. The results show these models achieve competitive or better task accuracy than standard neural models, provide meaningful concept semantics, support more effective test-time interventions, and scale to real-world conditions with incomplete supervision.
Deploying AI-powered systems requires trustworthy models supporting effective human interactions, going beyond raw prediction accuracy. Concept bottleneck models promote trustworthiness by conditioning classification tasks on an intermediate level of human-like concepts. This enables human interventions which can correct mispredicted concepts to improve the model's performance. However, existing concept bottleneck models are unable to find optimal compromises between high task accuracy, robust concept-based explanations, and effective interventions on concepts -- particularly in real-world conditions where complete and accurate concept supervisions are scarce. To address this, we propose Concept Embedding Models, a novel family of concept bottleneck models which goes beyond the current accuracy-vs-interpretability trade-off by learning interpretable high-dimensional concept representations. Our experiments demonstrate that Concept Embedding Models (1) attain better or competitive task accuracy w.r.t. standard neural models without concepts, (2) provide concept representations capturing meaningful semantics including and beyond their ground truth labels, (3) support test-time concept interventions whose effect in test accuracy surpasses that in standard concept bottleneck models, and (4) scale to real-world conditions where complete concept supervisions are scarce.