Leveraging the expertise of physicians to identify medically meaningful features in ‘counterfactual’ images produced via generative machine learning facilitates the auditing of the inference process of medical-image classifiers, as shown for dermatology images.
References
Gordon, R. Semin. Oncol. Nurs. 29, 160–169 (2013).
Zhang, X., Lin, D., Pforsich, H. & Lin, V. W. Hum. Resour. Health 18, 8 (2020).
Lorkowski, J. & Jugowicz, A. Adv. Exp. Med. Biol. 1324, 57–62 (2021).
Tsang, M. W. & Resneck, J. S. Jr J. Am. Acad. Dermatol. 55, 54–58 (2006).
Colin, J., Fel, T., Cadène, R. & Serre, T. Adv. Neural Inf. Process. Syst. 35, 2832–2845 (2022).
Adebayo, J. et al. Adv. Neural Inf. Process. Syst. 32, 9525–9536 (2018).
DeGrave, A. J., Janizek, J. D. & Lee, S.-I. Nat. Mach. Intell. 3, 610–619 (2021).
DeGrave, A. J., Cai, Z. R., Janizek, J. D., Daneshjou, R. & Lee, S.-I. Nat. Biomed. Eng. https://doi.org/10.1038/s41551-023-01160-9 (2023).
Wiens, J. et al. Nat. Med. 25, 1337–1340 (2019).
Lang, O. et al. Preprint at https://arxiv.org/abs/2306.00985 (2023).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Competing interests
O.L. and Y.L. are employees of Google and own Alphabet stock.
Rights and permissions
About this article
Cite this article
Lang, O., Traynis, I. & Liu, Y. Explaining counterfactual images. Nat. Biomed. Eng (2023). https://doi.org/10.1038/s41551-023-01164-5
Published:
DOI: https://doi.org/10.1038/s41551-023-01164-5
- Springer Nature Limited