Paper: Lexical Selection for Hybrid MT with Sequence Labeling

ACL ID W13-2815
Title Lexical Selection for Hybrid MT with Sequence Labeling
Venue Workshop on Hybrid Approaches to Translation
Session
Year 2013
Authors

We present initial work on an inex- pensive approach for building large- vocabulary lexical selection modules for hybrid RBMT systems by framing lexi- cal selection as a sequence labeling prob- lem. We submit that Maximum Entropy Markov Models (MEMMs) are a sensible formalism for this problem, due to their ability to take into account many features of the source text, and show how we can build a combination MEMM/HMM sys- tem that allows MT system implemen- tors flexibility regarding which words have their lexical choices modeled with classi- fiers. We present initial results showing successful use of this system both in trans- lating English to Spanish and Spanish to Guarani.