Semi-supervised Sequence Learning

TL;DR

This paper introduces two semi-supervised pretraining methods for recurrent neural networks, improving their stability and generalization in sequence tasks like text classification.

Contribution

It proposes two novel unsupervised pretraining approaches for LSTM networks, enhancing their performance in supervised sequence learning tasks.

Findings

Pretraining improves LSTM stability and generalization.

Pretrained LSTMs handle longer sequences effectively.

Enhanced performance on text classification benchmarks.

Abstract

We present two approaches that use unlabeled data to improve sequence learning with recurrent networks. The first approach is to predict what comes next in a sequence, which is a conventional language model in natural language processing. The second approach is to use a sequence autoencoder, which reads the input sequence into a vector and predicts the input sequence again. These two algorithms can be used as a "pretraining" step for a later supervised sequence learning algorithm. In other words, the parameters obtained from the unsupervised step can be used as a starting point for other supervised training models. In our experiments, we find that long short term memory recurrent networks after being pretrained with the two approaches are more stable and generalize better. With pretraining, we are able to train long short term memory recurrent networks up to a few hundred timesteps, thereby achieving strong performance in many text classification tasks, such as IMDB, DBpedia and 20 Newsgroups.