Learning Optimal Representations with the Decodable Information Bottleneck
arXiv:2009.12789
Abstract
We address the question of characterizing and finding optimal representations for supervised learning. Traditionally, this question has been tackled using the Information Bottleneck, which compresses the inputs while retaining information about the targets, in a decoder-agnostic fashion. In machine learning, however, our goal is not compression but rather generalization, which is intimately linked to the predictive family or decoder of interest (e.g. linear classifier). We propose the Decodable Information Bottleneck (DIB) that considers information retention and compression from the perspective of the desired predictive family. As a result, DIB gives rise to representations that are optimal in terms of expected test performance and can be estimated with guarantees. Empirically, we show that the framework can be used to enforce a small generalization gap on downstream classifiers and to predict the generalization ability of neural networks.
arXiv commentAccepted at NeurIPS 2020
References in corpus (27)
- PyTorch: An Imperative Style, High-Performance Deep Learning Library
- The information bottleneck method
- Learning deep representations by mutual information estimation and maximization
- Opening the Black Box of Deep Neural Networks via Information
- Estimating divergence functionals and the likelihood ratio by convex risk minimization
- Game theory, maximum entropy, minimum discrepancy and robust Bayesian decision theory
- Unrolled Generative Adversarial Networks
- Gradient-based Hyperparameter Optimization through Reversible Learning
- Regularizing Neural Networks by Penalizing Confident Output Distributions
- Sensitivity and Generalization in Neural Networks: an Empirical Study
- On Mutual Information Maximization for Representation Learning
- On the importance of single directions for generalization
- Learning Representations for Neural Network-Based Classification Using the Information Bottleneck Principle
- Path-SGD: Path-Normalized Optimization in Deep Neural Networks
- Nonlinear Information Bottleneck
- Predicting the Generalization Gap in Deep Networks with Margin Distributions
- Generalized Inner Loop Meta-Learning
- Norm-Based Capacity Control in Neural Networks
- Relevant sparse codes with variational information bottleneck
- A Minimax Approach to Supervised Learning
- Data-dependent Sample Complexity of Deep Neural Networks via Lipschitz Augmentation
- A Theory of Usable Information Under Computational Constraints
- Understanding Generalization through Visualizations
- Probabilistic supervised learning
- Minimal Achievable Sufficient Statistic Learning
- The Role of Information Complexity and Randomization in Representation Learning
- An Overview on Data Representation Learning: From Traditional Feature Learning to Recent Deep Learning