Learning Sparsity of Representations with Discrete Latent Variables

Abstract

Deep latent generative models have attracted increasing attention due to the capacity of combining the strengths of deep learning and probabilistic models in an elegant way. The data representations learned with the models are often continuous and dense. However in many applications, sparse representations are expected, such as learning sparse high dimensional embedding of data in an unsupervised setting, and learning multi-labels from thousands of candidate tags in a supervised setting. In some scenarios, there could be further restriction on degree of sparsity: the number of non-zero features of a representation cannot be larger than a pre-defined threshold $L_0$. In this paper we propose a sparse deep latent generative model SDLGM to explicitly model degree of sparsity and thus enable to learn the sparse structure of the data with the quantified sparsity constraint. The resulting sparsity of a representation is not fixed, but fits to the observation itself under the pre-defined restriction. In particular, we introduce to each observation $i$ an auxiliary random variable $L_i$, which models the sparsity of its representation. The sparse representations are then generated with a two-step sampling process via two Gumbel-Softmax distributions. For inference and learning, we develop an amortized variational method based on MC gradient estimator. The resulting sparse representations are differentiable with backpropagation. The experimental evaluation on multiple datasets for unsupervised and supervised learning problems shows the benefits of the proposed method.

References in corpus (18)

• Auto-Encoding Variational Bayes
• Categorical Reparameterization with Gumbel-Softmax
• Stochastic Backpropagation and Approximate Inference in Deep Generative Models
• Semi-Supervised Learning with Deep Generative Models
• Weight Uncertainty in Neural Networks
• The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables
• Unsupervised Anomaly Detection via Variational Auto-Encoder for Seasonal KPIs in Web Applications
• From Softmax to Sparsemax: A Sparse Model of Attention and Multi-Label Classification
• Variational Autoencoder for Deep Learning of Images, Labels and Captions
• Variational Dropout Sparsifies Deep Neural Networks
• Neural Variational Inference for Text Processing
• Fast Inference in Sparse Coding Algorithms with Applications to Object Recognition
• REBAR: Low-variance, unbiased gradient estimates for discrete latent variable models
• Deep AutoRegressive Networks
• Semantic Facial Expression Editing using Autoencoded Flow
• Training Sparse Neural Networks
• Efficient end-to-end learning for quantizable representations
• Minimizing FLOPs to Learn Efficient Sparse Representations