Lapshinova-Koltunski, E. and Santanu Pal. 2014. Comparability of Corpora in Human and Machine Translation. In Proceedings of BUCC, 7th Workshop on Building and Using Comparable Corpora. Building Resources for Machine Translation Research, (BUCC-2014), Reykjavik, May 27, 2014
Santanu Pal, Sudip Kumar Naskar and Sivaji Bandyopadhyay (2014) Word Alignment-Based Reordering of Source Chunks in PB-SMT. In Proceedings of the 9th International Conference on Language Resources and Evaluation (LREC-2014), Reykjavik, Iceland.
A alternative version of Moses formatted phrase table creation. The output of this program is the same format as Moses toolkit. The output phrase table is also accepted by the Moses decoder.
Developers can produce custom phrases (using their own algorithm) and append them to the phrase extracted file (extract.direct and extract.inv files)
along with the alignment as well as training file and the code produces Moses format phrase table.
An automatic statistical post editing system which extracts word alignments from METEOR based confusion network and it also gives Hybrid monolingual word alignments (HybridAlignment.java) (USAAR-SAPE: An English–Spanish Statistical Automatic Post-Editing System) . The repository also contains Java code which extract features from TER tool from a given MT output and the corresponding post editing version (runTer.java and FeatureExtraction.java). The repository also contain some evaluation code e.g.
A hybrid English-German Machine Translation system that participated at the SMT tool contest at ICON 2014. The hybrid system also extracts, named entities from POS tagged corpus and aligns them. It is also able to align verb chunks for a given POS tagged aligned corpus. Apart from this, it can also, identify MWEs using statistical method and align them.
