Pharaoh: A Beam Search Decoder for Phrase-Based Statistical Machine Translation Models

By Philipp Koehn. Conference of the Association for Machine Translation in the Americas, 2004

760 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/50b8e8d48f4973cdeefa835807b4e1a8ca65ced3

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Pivot language approach for phrase-based statistical machine translation (2007, methodology): “To perform phrase-based SMT, we use Koehn's training scripts1 and the Pharaoh decoder (Koehn, 2004).”
Deep Learning for Symbolic Mathematics (2019, methodology): “At inference, expressions are generated by a beam search (Koehn, 2004; Sutskever et al., 2014), with early stopping.”
Re-evaluating the Role of Bleu in Machine Translation Research (2006, positive): “We built the ﬁrst phrase-based SMT sys-tem with the complete set of Europarl data (14-15 million words per language), and optimized its feature functions using minimum error rate training in the standard way (Koehn, 2004).”
Paraphrasing with Bilingual Parallel Corpora (2005, neutral): “Koehn (2004), Tillmann (2003), and Vogel et al. (2003) describe various heuristics for extracting phrase alignments from the Viterbi word-level alignments that are estimated us-ing Brown et al. (1993)”
Learning Bilingual Lexicons from Monolingual Corpora (2008, background): “Current statistical machine translation systems use parallel corpora to induce translation correspondences, whether those correspondences be at the level of phrases (Koehn, 2004), treelets (Galley et al., 2006), or simply single words (Brown et al., 1994).”
A Hierarchical Phrase-Based Model for Statistical Machine Translation (2005, methodology): “When we run a phrase-based system, Pharaoh (Koehn et al., 2003; Koehn, 2004a), on this sentence (using the experimental setup described below), we get the following phrases with translations: where we have used subscripts to indicate the re-ordering of phrases.”
Statistical Significance Tests for Machine Translation Evaluation (2004, methodology): “We carry out experiments using a phrase-based statistical machine translation system [Koehn et al., 2003; Koehn, 2004].”
Word Sense Disambiguation Improves Statistical Machine Translation (2007, methodology): “Using the bootstrap-sampling test described in (Koehn, 2004b), the improvement is statistically signiﬁcant at p 0.05.”
(Meta-) Evaluation of Machine Translation (2007, neutral): “, 2006), which replaces the Pharaoh decoder (Koehn, 2004) • a training script to build models for Moses”
Forest Rescoring: Faster Decoding with Integrated Language Models (2007, methodology): “We set the decoder phrase-table limit to 100 as suggested in (Koehn, 2004) and the distortion limit to 4.”
Moses: Open Source Toolkit for Statistical Machine Translation (2007, background): “It features all the capabilities of the closed sourced Pharaoh decoder (Koehn 2004).”
Explainable Outfit Recommendation with Joint Outfit Matching and Comment Generation (2018, methodology): “During testing, for comment generation, we use beam search [51] to get better results.”
Neural Rating Regression with Abstractive Tips Generation for Recommendation (2017, methodology): “At the operational or testing stage, we use a beam search algorithm [13] for decoding and generating the best tips given a trained model.”
Improved Statistical Machine Translation Using Paraphrases (2006, methodology): “We used the Pharaoh beam-search decoder (Koehn, 2004) to produce the translations after all of the model parameters had been set.”
A Simple and Effective Hierarchical Phrase Reordering Model (2008, methodology): “Except for lexical reordering models, all other features are standard features implemented almost exactly as in Moses: four translation features (phrase-based translation probabilities and lexically-weighted probabilities), word penalty, phrase penalty, linear distortion, and language model score.”
A Survey of Paraphrasing and Textual Entailment Methods (2009, methodology): “Hence, heuristic search algorithms (e.g., based on beam search) are usually employed (Germann, Jahr, Knight, Marcu, & Yamada, 2001; Koehn, 2004).28 Assuming for simplicity that the individual deformations di(·) of D are mutually independent, P(F,D|N) can be computed as the product of the…”
Improving Arabic-Chinese Statistical Machine Translation using English as Pivot Language (2009, methodology): “These scores are computed over the grow-diag-final symmetrized alignment we use in our system (Koehn, 2004).”