Multiscale sequence modeling with a learned dictionary
Machine Learning
2017-07-06 v2 Machine Learning
Abstract
We propose a generalization of neural network sequence models. Instead of predicting one symbol at a time, our multi-scale model makes predictions over multiple, potentially overlapping multi-symbol tokens. A variation of the byte-pair encoding (BPE) compression algorithm is used to learn the dictionary of tokens that the model is trained with. When applied to language modelling, our model has the flexibility of character-level models while maintaining many of the performance benefits of word-level models. Our experiments show that this model performs better than a regular LSTM on language modeling tasks, especially for smaller models.
Cite
@article{arxiv.1707.00762,
title = {Multiscale sequence modeling with a learned dictionary},
author = {Bart van Merriënboer and Amartya Sanyal and Hugo Larochelle and Yoshua Bengio},
journal= {arXiv preprint arXiv:1707.00762},
year = {2017}
}