Báo cáo khoa học: "Cross-Language Document Summarization Based on Machine Translation Quality Prediction" pdf

... prediction and cross-language summarization, respectively. We discuss in Section 6 and conclude this paper in Section 7. 2 Related Work 2.1 Machine Translation Quality Prediction Machine translation ... Document summarization methods can be gener- ally categorized into extraction -based methods and abstraction -based methods. In this paper, we focus on extraction -ba...

Ngày tải lên: 07/03/2014, 22:20

... the functionality of summarization and content -based retrieval of tagged documents. This paper focuses on summarization based on this system. The main features of our summariza- tion method ... their documents as part of content authoring. This paper discusses au- tomatic text summarization based on GDA. Its main features are a domain/style-free algorithm and per- s...

Ngày tải lên: 31/03/2014, 04:20

... of weighted prediction model in this section is incrementally calculated by using only the informa- tion on the current state, thus the condition of state merge in Equation 2 remains unchanged. 5.3 ... their preconditions i < h, i = h and h < i do not hold at the same time. However, this algorithm faces a pred  -comp con- flict because both actions share the same precondi- tion h <...

Ngày tải lên: 16/03/2014, 19:20

... for the multi -document summarization task. 1 Introduction With the continuing growth of online text resources, document summarization has found wide-ranging applications in information retrieval and ... 2004)), non-negative matrix factorization (NMF) based methods (e.g., (Lee and Seung, 2001)), Conditional random field (CRF) based summarization (Shen et al., 2007), and LSA based...

Ngày tải lên: 23/03/2014, 17:20

... questions were designed to assess the content of the summary. We based our questions on the Responsive evaluation at DUC 2005; however, we were interested in a more spe- cific evaluation of the content ... worked on the questionnaire. Our questionnaire consisted of nine questions. The first five questions were the SEE linguistic well-formedness questions used at the 2005 Doc- ument Underst...

Ngày tải lên: 24/03/2014, 03:20

... is not only extraction but generation of ency- clopedic knowledge. Section 2 explains the overall design of our ency- clopedia generation system, and Section 3 elaborates on our organization model. ... and construc- tion domains (fields), respectively. To sum up, the organization module classifies term descriptions based on domains, for which we use do- main and description models. In Sec...

Ngày tải lên: 20/02/2014, 18:20

... combination of the weights of insertion, deletion and substi- tution. The relation is shown in equation (2), where ins, del and sub are the appearance times of insertions, deletions and substitutions, ... method are proposed in section 4. The experiments and the results are presented in section 5. The final section presents the conclusion and future work. 2 Related Work Conventional AES system...

Ngày tải lên: 07/03/2014, 18:20

... normalized current duration and the following one. In the above description, we assumed that the event of a syllable is only dependent on its observa- tions, and did not consider contextual effect. ... lexical stress based on the pronunciation dictionary. • Boundary information: This is a binary feature to indicate if there is a word boundary before the syllable. For lexical features,...

Ngày tải lên: 07/03/2014, 22:20

... Association for Computational Linguistics:shortpapers, pages 53–57, Portland, Oregon, June 19-24, 2011. c 2011 Association for Computational Linguistics Latent Class Transliteration based on Source ... lan- guages. Accurate transliteration is also the key to robust machine translation systems. Phonetic -based rewriting models (Knight and Jonathan, 1998) and spelling -based supervis...

Ngày tải lên: 07/03/2014, 22:20

... Contextual information: The simplest first-order word -based HMM with uniform initial conditions, achieves error reduction of 17.5% (78.2 – 82.01). (2) Initial conditions: Error reduc- tions in the range: ... number of emissions per sentence. However, we observe in our data that most of the words have only one or two possible segmentations, and most of the segmentations consist of at most one...

Ngày tải lên: 23/03/2014, 18:20