The CMU Machine Translation Systems at WMT 2014

Abstract

We describe the CMU systems submitted to the 2014 WMT shared translation task. We participated in two language pairs, German-English and Hindi-English. Our innovations include: a label coarsening scheme for syntactic tree-to-tree translation, a host of new discriminative features, several modules to create "synthetic translation options" that can generalize beyond what is directly observed in the training data, and a method of combining the output of multiple word aligners to uncover extra phrase pairs and grammar rules.

Findings of the 2014 Workshop on Statistical Machine Translation

“…AFRL, AFRL-PE Air Force Research Lab (Schwartz et al, 2014) CIMS University of Stuttgart / University of Munich (Cap et al, 2014) CMU Carnegie Mellon University (Matthews et al, 2014) CU-* Charles University, Prague (Tamchyna et al, 2014) DCU-FDA Dublin City University DCU-ICTCAS Dublin City University (Li et al, 2014b)…”

Detecting Cross-Lingual Semantic Divergence for Neural Machine Translation

“…Data Cleaning This line of work aims to remove noise, e.g., from alignment errors, based on scores from word alignment or language models (Okita et al, 2009;Jiang et al, 2010;Denkowski et al, 2012;Matthews et al, 2014). Cleaning training data in high-resource settings (Denkowski et al, 2012) and tuning data in lower resource settings (Matthews et al, 2014) has been shown to improve hierarchical phrase-based systems.…”

Findings of the 2015 Workshop on Statistical Machine Translation

“…AFRL, AFRL-PE Air Force Research Lab (Schwartz et al, 2014) CIMS University of Stuttgart / University of Munich CMU Carnegie Mellon University (Matthews et al, 2014) CU-* Charles University, Prague DCU-FDA Dublin City University DCU-ICTCAS Dublin City University (Li et al, 2014b) DCU-LINGO24 Dublin City University / Lingo24 (wu et al, 2014) EU-BRIDGE EU-BRIDGE Project KIT Karlsruhe Institute of Technology IIT-BOMBAY IIT Bombay (Dungarwal et al, 2014) IIIT-HYDERABAD IIIT Hyderabad…”