作者: Lucia Specia
DOI:
关键词: Machine translation 、 Information retrieval 、 Deliverable 、 User interface 、 Data mining 、 Software 、 Documentation 、 Variety (cybernetics) 、 Language model 、 Quality (business) 、 Computer science
摘要: We present an extended version of our open source framework for machine translation quality estimation, QUEST. The allows the extraction a large variety language- and system-dependent indicators from segments, their translations, external resources (i.e. target corpora, language models). It also provides learning algorithms to build estimation models. This new adds it more advanced, language-specific, system-related indicators. improvements over latest with respect efficiency user-friendliness. deliverable describes architecture framework, list additional features, documentation on how further extend add use models given dataset. Finally, presents web interface remote access facilitate by non-expert users.