Project overview: Of the many diverse world languages, very few are within reach of current natural language processing (NLP) and machine translation (MT) techniques. While mainstream approaches fail to generalize for most languages due to the lack of resources (e.g. text, annotations ...etc), our approach is designed to discover and leverage deep syntactic and semantic structures elicited from human experts.
Focus languages: Malagasy, Kinyarwanda, Swahili and Yoruba.
- Aug 2014: Ammar et al's paper on CRF autoencoders was selected for an oral presentation at NIPS 2014.
- July 2014: Lei et al's paper on low rank tensors for dependency parsing won the best student paper award at ACL 2014.
- May 2014: A paper was accepted at the SemEval Workshop 2014.
- May 2014: Four publications accepted at ACL 2014 on low rank tensors, using arbitrary global features for parsing, super taggers, and AMR parsers.
- January 2014: Two publications accepted at LREC 2014 on adjectives and definiteness.
- December 2013: A paper on spoken language translation accepted at EACL2014.
- September 2013: Two publications accepted at EMNLP 2013 on translation into morphologically rich languages and dependency-based decipherment.
- August 2013: An improved Kinyarwanda morphological analyzer has been released!
- July 2013: A paper on synthetic translation options accepted at WMT 2013.
- June 2013: A paper on quasi-synchronous tree-to-tree MT was accepted for publication in the Computational Linguistics journal.
- April 2013: Three publications accepted at ACL 2013 on POS tagging, transfer learning of grammars and parsing graphs.
- March 2013: Cross-site technical meeting at Pittsburgh, PA to follow up on collaboration projects.
- February 2013: Four publications accepted at NAACL 2013 on word alignment, POS tagging, large-scale discriminative training, and language modeling.