Learning to Skim Text
This addresses the slow processing of long documents in NLP tasks, offering a practical speed-up for applications like document classification and Q&A, though it is an incremental improvement over existing LSTM methods.
The paper tackles the inefficiency of recurrent neural networks in processing long texts by introducing a model that learns to skip irrelevant information, achieving up to 6 times faster processing while maintaining or improving accuracy across tasks like sentiment analysis and question answering.
Recurrent Neural Networks are showing much promise in many sub-areas of natural language processing, ranging from document classification to machine translation to automatic question answering. Despite their promise, many recurrent models have to read the whole text word by word, making it slow to handle long documents. For example, it is difficult to use a recurrent network to read a book and answer questions about it. In this paper, we present an approach of reading text while skipping irrelevant information if needed. The underlying model is a recurrent network that learns how far to jump after reading a few words of the input text. We employ a standard policy gradient method to train the model to make discrete jumping decisions. In our benchmarks on four different tasks, including number prediction, sentiment analysis, news article classification and automatic Q\&A, our proposed model, a modified LSTM with jumping, is up to 6 times faster than the standard sequential LSTM, while maintaining the same or even better accuracy.