TAILIEUCHUNG - Báo cáo khoa học: "Topic Models for Dynamic Translation Model Adaptation"

We propose an approach that biases machine translation systems toward relevant translations based on topic-specific contexts, where topics are induced in an unsupervised way using topic models; this can be thought of as inducing subcorpora for adaptation without any human annotation. We use these topic distributions to compute topic-dependent lexical weighting probabilities and directly incorporate them into our translation model as features. | Topic Models for Dynamic Translation Model Adaptation Vladimir Eidelman Computer Science and UMIACS University of Maryland College Park MD vlad@ Jordan Boyd-Graber iSchool and UMIACS University of Maryland College Park MD jbg@ Philip Resnik Linguistics and UMIACS University of Maryland College Park MD resnik@ Abstract We propose an approach that biases machine translation systems toward relevant translations based on topic-specific contexts where topics are induced in an unsupervised way using topic models this can be thought of as inducing subcorpora for adaptation without any human annotation. We use these topic distributions to compute topic-dependent lexical weighting probabilities and directly incorporate them into our translation model as features. Conditioning lexical probabilities on the topic biases translations toward topicrelevant output resulting in significant improvements of up to 1 BLEU and 3 TER on Chinese to English translation over a strong baseline. 1 Introduction The performance of a statistical machine translation SMT system on a translation task depends largely on the suitability of the available parallel training data. Domains . newswire vs. blogs may vary widely in their lexical choices and stylistic preferences and what may be preferable in a general setting or in one domain is not necessarily preferable in another domain. Indeed sometimes the domain can change the meaning of a phrase entirely. In a food related context the Chinese sentence W M ÍM fensi henduo would mean They have a lot of vermicelli however in an informal Internet conversation this sentence would mean They have a lot of fans . Without the broader context it is impossible to determine the correct translation in otherwise identical sentences. 115 This problem has led to a substantial amount of recent work in trying to bias or adapt the translation model TM toward particular domains of interest Axelrod et al. 2011 Foster et al. 2010 .

• Báo cáo khoa học
• Topic Models
• Dynamic Translation Model Adaptation
• Vladimir Eidelman
• báo cáo khoa học
• báo cáo ngôn ngữ
• ngôn ngữ tự nhiên
• BMC Bioinformatics
• Dynamic topic model
• Times series gene expression
• Topic modeling
• Latent dirichlet model
• Authorship Attribution
• Author aware Topic Models
• Yanir Seroussi
• Word Sense Disambiguation
• Token based Idiom Detection
• PCFGs Topic Models
• Adaptor Grammars and Learning Topical Collocations
• the Structure of Proper Names
• Automatic Labelling
• Jey Han Lau
• Identifying Word Translations
• Comparable Corpora
• Latent Topic Models
• Employing Topic Models
• Pattern based Semantic Class Discovery
• Mingjie Zhu
• xử lý ngôn ngữ tự nhiên
• Multi Document Summarization
• Sentence based Topic Models
• Dingding Wang
• Unsupervised Topic Identification
• Integrating Linguistic
• Visual Information Based on Hidden Markov Models
• mô hình ngôn ngữ
• Bilingual Topic AdMixture Models
• Word Alignment
• Bing Zhao
• Incorporating topic information
• sentiment analysis models
• Tony Mullen
• Statistical Models
• Topic Segmentation
• Jeffrey C
• Reynar
• Incorporating Lexical Priors
• Jagadeesh Jagarlamudi
• scientific reports
• model language
• process natural language
• Posterior inference
• Online maximum a posteriori estimation
• Large scale learning
• Streaming text collections
• Financial modeling
• Lecture Financial modeling
• Asset expected return
• Accrued interest
• Yield curve models
• Forward rates
• Financial statement simulation
• Cash flow
• Cash flow models
• Financial simulations
• Retirement ratios
• Changing asset allocation
• Sequence motifs
• Transcription factor binding sites
• Molecular biology
• DNA sequences
• User defined function
• Calculating cost of capital
• Portfolio models
• Estimating asset risk
• Asset risk
• Portfolio risk
• Computing portfolio risk