arxivst stuff from arxiv that you should probably bookmark

Neural Machine Translation via Binary Code Prediction

Abstract · Apr 23, 2017 12:38 ·

nmt japan vocabulary gpus batched softmax computation memory output cs-cl

Arxiv Abstract

  • Yusuke Oda
  • Philip Arthur
  • Graham Neubig
  • Koichiro Yoshino
  • Satoshi Nakamura

In this paper, we propose a new method for calculating the output layer in neural machine translation systems. The method is based on predicting a binary code for each word and can reduce computation time/memory requirements of the output layer to be logarithmic in vocabulary size in the best case. In addition, we also introduce two advanced approaches to improve the robustness of the proposed model: using error-correcting codes and combining softmax and binary codes. Experiments on two English-Japanese bidirectional translation tasks show proposed models achieve BLEU scores that approach the softmax, while reducing memory usage to the order of less than 110 and improving decoding speed on CPUs by x5 to x10.

Read the paper (pdf) »