Remix: Rebalanced Mixup

By Hsin-Ping Chou, Shih-Chieh Chang, Jia-Yu Pan, Wei Wei, Da-Cheng Juan. ECCV Workshops, 2020

304 citations recorded by the source corpus.

Deep image classifiers often perform poorly when training data are heavily class-imbalanced. In this work, we propose a new regularization technique, Remix, that relaxes Mixup's formulation and enables the mixing factors of features and labels to be disentangled. Specifically, when mixing two samples, while features are mixed in the same fashion as Mixup, Remix assigns the label in favor of the minority class by providing a disproportionately higher weight to the minority class. By doing so, the classifier learns to push the decision boundaries towards the majority classes and balance the generalization error between majority and minority classes. We have studied the state-of-the art regularization techniques such as Mixup, Manifold Mixup and CutMix under class-imbalanced regime, and shown that the proposed Remix significantly outperforms these state-of-the-arts and several re-weighting and re-sampling techniques, on the imbalanced datasets constructed by CIFAR-10, CIFAR-100, and CINIC-10. We have also evaluated Remix on a real-world large-scale imbalanced dataset, iNaturalist 2018. The experimental results confirmed that Remix provides consistent and significant improvements over the previous methods.

Subfields: Artificial Intelligence, Radiology, Nuclear Medicine and Imaging

View Paper: https://www.semanticscholar.org/paper/a50b451538891b0c3a9871ea4e213147d139a8d5

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Jiashi Feng
2. Shuicheng Yan
3. Jiaya Jia
4. Shu Liu
5. Yifan Zhang
6. Bingyi Kang
7. Jiequan Cui
8. Bryan Hooi

In Radiology, Nuclear Medicine and Imaging

1. Yifan Zhang
2. Jiashi Feng
3. Shuicheng Yan
4. Jiequan Cui
5. Jiaya Jia
6. Bingyi Kang
7. Shu Liu
8. Zhisheng Zhong

Citations in the corpus

• Improving Calibration for Long-Tailed Recognition (2020, methodology): “…model capacity, normalization, and regularization have strong effect on network calibration. mixup [37] is a regularization technique to train with interpolation of input and labels. mixup inspires follow-up of manifold mixup [32], Cut-Mix [36], and Remix [5] that have shown significant improvement.”
• Deep Long-Tailed Learning: A Survey (2021, methodology): “[160] H.-P. Chou, S.-C. Chang, J.-Y. Pan, W. Wei, and D.-C. Juan, “Remix: Rebalanced mixup,” in Proc.”