Deep Deterministic Information Bottleneck with Matrix-based Entropy Functional
This addresses the problem of enhancing neural network reliability for machine learning practitioners, but it appears incremental as it builds on the established information bottleneck principle with a new functional.
The paper tackled the problem of improving generalization and robustness in neural networks by introducing the Deep Deterministic Information Bottleneck (DIB), which uses a matrix-based entropy functional to parameterize the information bottleneck principle without variational inference. The result showed that DIB outperformed variational and other regularization methods in generalization performance and robustness to adversarial attacks, though no concrete numbers were provided.
We introduce the matrix-based Renyi's $α$-order entropy functional to parameterize Tishby et al. information bottleneck (IB) principle with a neural network. We term our methodology Deep Deterministic Information Bottleneck (DIB), as it avoids variational inference and distribution assumption. We show that deep neural networks trained with DIB outperform the variational objective counterpart and those that are trained with other forms of regularization, in terms of generalization performance and robustness to adversarial attack.Code available at https://github.com/yuxi120407/DIB