Arya, S., Rao, S., Böhle, M., & Schiele, B. (2024). B-cosification: Transforming Deep Neural Networks to be Inherently Interpretable [Conference paper]. Advances in Neural Information Processing Systems