Contrastive Learning in Medical Imaging - A metric learning approach
Pietro Gori
Abstract
Contrastive Learning (CL) is a paradigm designed for self-supervised representation learning which has been applied to unsupervised, weakly supervised and supervised problems. The objective in CL is to estimate a parametric mapping function that maps positive samples (semantically similar) close together in the representation space and negative samples (semantically dissimilar) far away from each other. In general, positive samples can be defined in different ways depending on the problem: transformations (i.e., augmentations) of the same image (unsupervised setting), samples belonging to the same class (supervised) or with similar image attributes (weakly-supervised). The definition of negative samples varies accordingly. In this talk, we will show how a metric learning approach for CL allows us to: 1- better formalize well-known contrastive losses, such as InfoNCE and SupCon, 2- derive new losses for unsupervised, supervised, and weakly supervised problems, both classification and regression, and 3- propose new regularization terms for debiasing. Furthermore, leveraging the proposed metric learning approach and kernel theory, we will describe a novel loss, called decoupled uniformity, that allows the integration of prior knowledge, given either by generative models or weak attributes, and removes the positive-negative coupling problem, as in the InfoNCE loss. We validate the usefulness of the proposed losses on standard vision datasets and medical imaging data.
Bio
Pietro Gori (PhD, HDR) is Professor in Artificial Intelligence and Medical Imaging at Télécom Paris (IPParis) in the IMAGES group. He completed his PhD with Inria at the ARAMIS Lab in Paris and worked as a post-doctoral researcher at Neurospin (CEA). Previous to that, he obtained a MSc in Mathematical Modeling and Computation from the DTU in Copenhagen and a MSc in Biomedical Engineering from the University of Padova. He participated to the development of the open source software suite deformetrica for statistical shape analysis and of the software platform Clinica for clinical neuroimaging studies. His research interests lie primarily in the fields of machine learning, AI, representation learning, medical imaging and computational anatomy. He has more than 60 publications in international peer-reviewed journals or conferences, 3 patents and had/has the pleasure to work with 15 Master students, 18 PhD students and 3 post-docs. He is one of the co-founders of the Start-Up Replico.
References
-
Dufumier, B., Gori, P., Petiton, S., Louiset, R., Mangin, J.-F., Grigis, A., & Duchesnay, E. (2024). Exploring the potential of representation and transfer learning for anatomical neuroimaging: Application to psychiatry. NeuroImage.
-
Louiset, R., Duchesnay, E., Grigis, A., & Gori, P. (2024). Separating common from salient patterns with Contrastive Representation Learning. In Proceedings of the International Conference on Learning Representations (ICLR).
-
Barbano, C. A., Dufumier, B., Tartaglione, E., Grangetto, M., & Gori, P. (2023). Unbiased Supervised Contrastive Learning. In Proceedings of the International Conference on Learning Representations (ICLR).
-
Dufumier, B., Barbano, C. A., Louiset, R., Duchesnay, E., & Gori, P. (2023). Integrating Prior Knowledge in Contrastive Learning with Kernel. In Proceedings of the International Conference on Machine Learning (ICML).
-
Barbano, C. A., Dufumier, B., Duchesnay, E., Grangetto, M., & Gori, P. (2023). Contrastive Learning for Regression in Multi-Site Brain Age Prediction. In Proceedings of the IEEE International Symposium on Biomedical Imaging (ISBI).
-
Dufumier, B., Gori, P., Victor, J., Grigis, A., & Duchesnay, E. (2021). Contrastive Learning with Continuous Proxy Meta-Data for 3D MRI Classification. In Proceedings of the Medical Image Computing and Computer-Assisted Intervention Conference (MICCAI).