Learning Diverse and Discriminative Representations via the Principle of Maximal Coding Rate Reduction
By Yaodong Yu, Kwan Ho Ryan Chan, Chong You, Chaobing Song, Yi Ma. 2020
251 citations recorded by the source corpus.
To learn intrinsic low-dimensional structures from high-dimensional data that most discriminate between classes, we propose the principle of Maximal Coding Rate Reduction ($\text{MCR}^2$), an information-theoretic measure that maximizes the coding rate difference between the whole dataset and the sum of each individual class. We clarify its relationships with most existing frameworks such as cross-entropy, information bottleneck, information gain, contractive and contrastive learning, and provide theoretical guarantees for learning diverse and discriminative features. The coding rate can be accurately computed from finite samples of degenerate subspace-like distributions and can learn intrinsic representations in supervised, self-supervised, and unsupervised settings in a unified manner. Empirically, the representations learned using this principle alone are significantly more robust to label corruptions in classification than those using cross-entropy, and can lead to state-of-the-art results in clustering mixed data from self-learned invariant features.
Subfields: Artificial Intelligence
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
Citations in the corpus
- Deep Networks from the Principle of Rate Reduction (2020, methodology): “Notice that these are the cases to which the existing theory about MCR2 (Yu et al., 2020) no longer applies.”
- Incremental Learning via Rate Reduction (2021, background): “In [25], it is demonstrated empirically that maximizing ∆R(Z) enforces each latent class distribution to be a low-dimensional subspace-like distribution of approximately d k dimension with class balance i.”