Oracle decoding as a new way to analyze phrase-based machine translation - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Machine Translation Année : 2013

Oracle decoding as a new way to analyze phrase-based machine translation

Guillaume Wisniewski
François Yvon

Résumé

Extant Statistical Machine Translation systems are very complex pieces of software, which embed multiple layers of heuristics and encompass very large numbers of numerical parameters. As a result, it is difficult to analyze output translations and there is a real need for tools that could help developers to better understand the various causes of errors. In this study, we make a step in that direction and present an attempt to evaluate the quality of the phrase-based translation model. In order to identify those translation errors that stem from deficiencies in the phrase table, we propose to compute the oracle BLEU-4 score, that is the best score that a system based on this phrase table can achieve on a reference corpus. By casting the computation of the oracle BLEU-1 as an Integer Linear Programming problem, we show that it is possible to efficiently compute accurate upper-bounds of this score, and report measures performed on several standard benchmarks. Various other applications of these oracle decoding techniques are also reported and discussed.
Fichier non déposé

Dates et versions

hal-01908378 , version 1 (30-10-2018)

Identifiants

  • HAL Id : hal-01908378 , version 1

Citer

Guillaume Wisniewski, François Yvon. Oracle decoding as a new way to analyze phrase-based machine translation. Machine Translation, 2013, 27, pp.115-138. ⟨hal-01908378⟩
22 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More