Publication

Sequence Level Training with Recurrent Neural Networks

ICLR 2016


Abstract

Many natural language processing applications use language models to generate text. These models are typically trained to predict the next word in a sequence, given the previous words and some context such as an image. However, at test time the model is expected to generate the entire sequence from scratch. This discrepancy makes generation brittle, as errors may accumulate along the way. We address this issue by proposing a novel sequence level training algorithm that directly optimizes the metric used at test time, such as BLEU or ROUGE. On three different tasks, our approach outperforms several strong baselines for greedy generation. The method is also competitive when these baselines employ beam search, while being several times faster.

Related Publications

All Publications

December 2, 2016

MIXER

No Authors Listed

EACL - April 18, 2021

Co-evolution of language and agents in referential games

Gautier Dagan, Dieuwke Hupkes, Elia Bruni

PPSN - September 2, 2020

Variance Reduction for Better Sampling in Continuous Domains

Laurent Meunier, Carola Doerr, Jeremy Rapin, Olivier Teytaud

To help personalize content, tailor and measure ads, and provide a safer experience, we use cookies. By clicking or navigating the site, you agree to allow our collection of information on and off Facebook through cookies. Learn more, including about available controls: Cookies Policy