KATE: K-Competitive Autoencoder for Text

TL;DR

KATE introduces a k-competitive autoencoder that enhances text representation learning by encouraging neuron specialization, outperforming traditional autoencoders and other models in various downstream NLP tasks.

Contribution

The paper proposes the KATE model, a novel k-competitive autoencoder that improves text representation learning through neuron competition, leading to better performance than existing autoencoders and generative models.

Findings

KATE learns more meaningful text representations than traditional autoencoders.

KATE outperforms deep generative and probabilistic models in downstream tasks.

Neuron competition in KATE enhances specialization and model effectiveness.

Abstract

Autoencoders have been successful in learning meaningful representations from image datasets. However, their performance on text datasets has not been widely studied. Traditional autoencoders tend to learn possibly trivial representations of text documents due to their confounding properties such as high-dimensionality, sparsity and power-law word distributions. In this paper, we propose a novel k-competitive autoencoder, called KATE, for text documents. Due to the competition between the neurons in the hidden layer, each neuron becomes specialized in recognizing specific data patterns, and overall the model can learn meaningful representations of textual data. A comprehensive set of experiments show that KATE can learn better representations than traditional autoencoders including denoising, contractive, variational, and k-sparse autoencoders. Our model also outperforms deep generative models, probabilistic topic models, and even word representation models (e.g., Word2Vec) in terms of several downstream tasks such as document classification, regression, and retrieval.