Method and system for improving machine translation accuracy using translation memory
作者： Dennis D. King , Richard J. Redpath
DOI:
关键词: Speech recognition 、 Artificial intelligence 、 Translation (geometry) 、 Natural language processing 、 Matching (graph theory) 、 Translation memory 、 Value (computer science) 、 Machine translation 、 Sentence 、 Rule-based machine translation 、 Computer science
摘要: A combination of Machine Translation (MT) techniques are used with a ‘Hybrid’ Memory (HTM)database. The HTM database stores malformed translated sentences matched their correct translations. Each sentence that is MT analyzed by the algorithms to determine value for accuracy translation. If greater than or equal preset tolerance level, considered be sufficiently well. Therefore, no need checked all sentences. However, if below tolerance, an same bad translation and found, replaced matching
lens.org 参考文章(4)
Robert Todd Ward, Kishore Ananda Papineni, Salim Estephan Roukos, Statistical translation system with features based on phrases or groups of words ,(1997)
Vincent Joseph Della Pietra, Frederick Jelinek, Jennifer Ceil Lai, John Cocke, Stephen Andrew Della Pietra, Robert Leroy Mercer, Peter Fitzhugh Brown, Method and system for natural language translation ,(1992)
Timothy Francis O'Donoghue, Method, an apparatus, a system, a storage device, and a computer readable medium using a bilingual database including aligned corpora ,(1994)
Akira Kumano, Hideki Hirakawa, Dictionary creation supporting system. ,(1994)