On Diversity in Discriminative Neural Networks
This work addresses the need for improved diversity in neural networks for machine learning practitioners, offering incremental advancements by combining known and original principles.
The paper tackled the problem of underemphasized diversity in neural networks by proposing a new architecture based on diversity principles, achieving a record self-supervised learning accuracy of 99.57% on MNIST and a promising semi-supervised learning accuracy of 94.21% on CIFAR-10 with only 25 labels per class.
Diversity is a concept of prime importance in almost all disciplines based on information processing. In telecommunications, for example, spatial, temporal, and frequency diversity, as well as redundant coding, are fundamental concepts that have enabled the design of extremely efficient systems. In machine learning, in particular with neural networks, diversity is not always a concept that is emphasized or at least clearly identified. This paper proposes a neural network architecture that builds upon various diversity principles, some of them already known, others more original. Our architecture obtains remarkable results, with a record self-supervised learning accuracy of 99. 57% in MNIST, and a top tier promising semi-supervised learning accuracy of 94.21% in CIFAR-10 using only 25 labels per class.