SeaRNN: Training RNNs with Global-Local Losses - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2018

SeaRNN: Training RNNs with Global-Local Losses

Abstract

We propose SEARNN, a novel training algorithm for recurrent neural networks (RNNs) inspired by the "learning to search" (L2S) approach to structured prediction. RNNs have been widely successful in structured prediction applications such as machine translation or parsing, and are commonly trained using maximum likelihood estimation (MLE). Unfortunately, this training loss is not always an appropriate surrogate for the test error: by only maximizing the ground truth probability, it fails to exploit the wealth of information offered by structured losses. Further, it introduces discrepancies between training and predicting (such as exposure bias) that may hurt test performance. Instead, SEARNN leverages test-alike search space exploration to introduce global-local losses that are closer to the test error. We first demonstrate improved performance over MLE on two different tasks: OCR and spelling correction. Then, we propose a subsampling strategy to enable SEARNN to scale to large vocabulary sizes. This allows us to validate the benefits of our approach on a machine translation task.
Fichier principal
Vignette du fichier
15c49b74558c239a4dbfa438d3abe751a51ac298.pdf (350.23 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01950555 , version 1 (10-12-2018)

Identifiers

  • HAL Id : hal-01950555 , version 1

Cite

Rémi Leblond, Jean-Baptiste Alayrac, Anton Osokin, Simon Lacoste-Julien. SeaRNN: Training RNNs with Global-Local Losses. ICLR 2018 : 6th International Conference on Learning Representations, Apr 2018, Vancouver, Canada. ⟨hal-01950555⟩
80 View
73 Download

Share

Gmail Facebook X LinkedIn More