
A new technique transforms any computer vision model into one that can explain its predictions using a set of concepts a human could understand. The method generates more appropriate concepts that boost the accuracy of the model.
No discussion yet. Be the first to share your thoughts!