Beyond Explanation: Debugging Medical Imaging Models via Concept Intervention

Key points

• Medical imaging models often operate as black boxes, limiting interpretability and systematic debugging.1
• By aligning a single-modality encoder to BioMedCLIP, we construct a Concept Bottleneck Model (CBM) that enables concept-level interventions.1
• Results demonstrate that concept intervention enables reliable model diagnosis while maintaining, and occasionally improving predictive performance via guided fine-tuning.1
• Our findings highlight the practical value of this framework for controlled, interpretable refinement of clinical deep learning models.1

Sources (1)

• [1]Beyond Explanation: Debugging Medical Imaging Models via Concept Intervention
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 07:31 PM

  We introduce an easy-to-use, plug-and-play framework for concept-based interpretation and model refinement.

  Medical imaging models often operate as black boxes, limiting interpretability and systematic debugging.

Extractive summary: sentences quoted from the sources.