Learning representations for individual instances when only bag-level labels are available is a fundamental challenge in multiple instance learning (MIL). Recent works have shown promising results using contrastive self-supervised learning (CSSL), which learns to push apart representations corresponding to two different randomly-selected instances. Unfortunately, in real-world applications such as medical image classification, there is often class imbalance, so randomly-selected instances mostly belong to the same majority class, which precludes CSSL from learning inter-class differences. To address this issue, we propose a novel framework, Iterative Self-paced Supervised Contrastive Learning for MIL Representations (ItS2CLR), which improves the learned representation by exploiting instance-level pseudo labels derived from the bag-level labels. The framework employs a novel self-paced sampling strategy to ensure the accuracy of pseudo labels. We evaluate ItS2CLR on three medical datasets, showing that it improves the quality of instance-level pseudo labels and representations, and outperforms existing MIL methods in terms of both bag and instance level accuracy. Code is available at https://github.com/Kangningthu/ItS2CLR
@article{arxiv.2210.09452,
title = {Multiple Instance Learning via Iterative Self-Paced Supervised Contrastive Learning},
author = {Kangning Liu and Weicheng Zhu and Yiqiu Shen and Sheng Liu and Narges Razavian and Krzysztof J. Geras and Carlos Fernandez-Granda},
journal= {arXiv preprint arXiv:2210.09452},
year = {2023}
}
Comments
CVPR 2023 camera-ready version. The first two authors contribute equally. The last two authors are joint last authors