Deep learning and the information bottleneck principle
Naftali Tishby,Noga Zaslavsky +1 more
- pp 1-5
TLDR
It is argued that both the optimal architecture, number of layers and features/connections at each layer, are related to the bifurcation points of the information bottleneck tradeoff, namely, relevant compression of the input layer with respect to the output layer.Abstract:
Deep Neural Networks (DNNs) are analyzed via the theoretical framework of the information bottleneck (IB) principle. We first show that any DNN can be quantified by the mutual information between the layers and the input and output variables. Using this representation we can calculate the optimal information theoretic limits of the DNN and obtain finite sample generalization bounds. The advantage of getting closer to the theoretical limit is quantifiable both by the generalization bound and by the network's simplicity. We argue that both the optimal architecture, number of layers and features/connections at each layer, are related to the bifurcation points of the information bottleneck tradeoff, namely, relevant compression of the input layer with respect to the output layer. The hierarchical representations at the layered network naturally correspond to the structural phase transitions along the information curve. We believe that this new insight can lead to new optimality bounds and deep learning algorithms.read more
Citations
More filters
Journal ArticleDOI
Sufficient Dimension Reduction: An Information-Theoretic Viewpoint
TL;DR: This paper developed an approach to interpret sufficient dimension reduction (SDR) techniques using information theory, which leads to a more assumption-lean understanding of what SDR methods do and also allows for some connections to results in the information theory literature.
Posted Content
On Data-Augmentation and Consistency-Based Semi-Supervised Learning
Atin Ghosh,Alexandre H. Thiery +1 more
TL;DR: In this article, the authors propose an extension of the Hidden Manifold Model that naturally incorporates data augmentation schemes and offer a framework for understanding and experimenting with semi-supervised learning methods.
Journal ArticleDOI
Partial Information Decomposition Reveals the Structure of Neural Representations
TL;DR: In this article , the authors introduce the measure of representational complexity, which quantifies the difference between accessing information spread across multiple neurons in neural networks, and show that this complexity is directly computable for smaller layers.
Posted ContentDOI
Single-cell membrane potential fluctuations evince network scale-freeness and quasicriticality
TL;DR: The results demonstrate appropriation of the brain's own subsampling method (convergence of synaptic inputs), while extending the range of fundamental evidence for critical branching in neural systems from the previously observed mesoscale to the microscale, namely, membrane potential fluctuations.
Journal ArticleDOI
Mutual Information Based Learning Rate Decay for Stochastic Gradient Descent Training of Deep Neural Networks.
TL;DR: This paper demonstrates a novel approach to training deep neural networks using a Mutual Information (MI)-driven, decaying Learning Rate (LR), Stochastic Gradient Descent (SGD) algorithm, and demonstrates the feasibility of the metric and approach.
References
More filters
Proceedings Article
ImageNet Classification with Deep Convolutional Neural Networks
TL;DR: The state-of-the-art performance of CNNs was achieved by Deep Convolutional Neural Networks (DCNNs) as discussed by the authors, which consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax.
Book
Elements of information theory
Thomas M. Cover,Joy A. Thomas +1 more
TL;DR: The author examines the role of entropy, inequality, and randomness in the design of codes and the construction of codes in the rapidly changing environment.
Journal ArticleDOI
Reducing the Dimensionality of Data with Neural Networks
TL;DR: In this article, an effective way of initializing the weights that allows deep autoencoder networks to learn low-dimensional codes that work much better than principal components analysis as a tool to reduce the dimensionality of data is described.
Journal ArticleDOI
Representation Learning: A Review and New Perspectives
TL;DR: Recent work in the area of unsupervised feature learning and deep learning is reviewed, covering advances in probabilistic models, autoencoders, manifold learning, and deep networks.
Book
Learning Deep Architectures for AI
TL;DR: The motivations and principles regarding learning algorithms for deep architectures, in particular those exploiting as building blocks unsupervised learning of single-layer modelssuch as Restricted Boltzmann Machines, used to construct deeper models such as Deep Belief Networks are discussed.