Standard Statistical Machine Translation Research Articles

FaDA: Fast Document Aligner using Word Embedding

Abstract FaDA is a free/open-source tool for aligning multilingual documents. It employs a novel crosslingual information retrieval (CLIR)-based document-alignment algorithm involving the distances between embedded word vectors in combination with the word overlap between the source-language and the target-language documents. In this approach, we initially construct a pseudo-query from a source-language document. We then represent the target-language documents and the pseudo-query as word vectors to find the average similarity measure between them. This word vector-based similarity measure is then combined with the term overlap-based similarity. Our initial experiments show that s standard Statistical Machine Translation (SMT)- based approach is outperformed by our CLIR-based approach in finding the correct alignment pairs. In addition to this, subsequent experiments with the word vector-based method show further improvements in the performance of the system.

Learning Machine Translation. * Cyril Goutte, Nicola Cancedda, Marc Dymetman, and George Foster

Learning Machine Translation is a collection of papers on using machine learning in machine translation (MT) from the workshop on Machine Learning for Multilingual Information Access, organized at the Neural Information Processing Systems (NIPS) 19 conference in December 2006. It is not a book for readers who would like to be initiated in MT, although it starts off with an introduction to statistical machine translation (SMT) by the editors themselves. Its description of the SMT paradigm is very technical and has clearly been written from a computer scientist’s viewpoint. A much clearer introduction to SMT, including examples and instructions on how to build your own MT system, is given by Koehn (2010). The book consists of two main parts. The first deals with enabling technologies—technologies that solve problems which are not exclusive to MT, but very relevant. Each chapter in this part applies machine learning to a different problem. The second part describes research into alternative solutions, involving machine learning, for certain components of standard SMT systems, trying to break through the current ceiling of translation quality that can be obtained using state-of-the-art phrase-based SMT systems.