multi rate hmms for word alignment

• DOI: 10.1184/r1/6287720
• Publication Date: 2013-08-01
• Authors: Eyigoz, Elif; Gildea, Daniel; Oflazer, Kemal
• Keywords: 200402 Computational Linguistics; Applied Computer Science; FOS: Languages and literature

We apply multi-rate HMMs, a tree structured HMM model, to the word-alignment problem. Multi-rate HMMs allow us to model reordering at both the morpheme level and the word level in a hierarchical fashion. This approach leads to better machine translation results than a morpheme aware model that does not explicitly model morpheme reordering.