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
Posted Content
Lattice Representation Learning
TL;DR: This article introduces theory and algorithms for learning discrete representations that take on a lattice that is embedded in an Euclidean space, including a new mathematical result linking expressions used during training and inference time and experimental validation on two popular datasets.
Journal ArticleDOI
Cross-Domain Gated Learning for Domain Generalization
TL;DR: The proposed CDG training strategy can excellently enforce the network to exploit the intrinsic features of objects from the multi-domain data, and achieve a new state-of-the-art domain generalization performance on these benchmarks.
Proceedings ArticleDOI
Dynamic Narrowing of VAE Bottlenecks Using GECO and L 0 Regularization
TL;DR: In this article, a technique to shrink the latent space dimensionality of VAEs automatically and on-the-fty during training using Generalized ELBO with Constrained Optimization (GECO) and the $L_{0}$ -Augment-REINFORcE-Merge (ARM) gradient estimator is presented.
Proceedings ArticleDOI
Learning photonic neural network initialization for noise-aware end-to-end fiber transmission
Manos Kirtas,N. Passalis,George Mourgias-Alexandris,George Dabos,Nikos Pleros,Anastasios Tefas +5 more
TL;DR: A trainable data-driven noise-aware initialization method oriented to easily saturated activation functions, such as those typically used in optical neurons on the transmitter and receiver side of a noisy IM/DD system intercepted by a noisy channel are presented.
Posted Content
A Generalization Theory based on Independent and Task-Identically Distributed Assumption.
TL;DR: A new Independent and Task-Identically Distributed (ITID) assumption is proposed, to consider the task properties into the data generating process and the derived generalization bound based on the ITID assumption identifies the significance of hypothesis invariance in guaranteeing generalization performance.
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.