Báo cáo khoa học: "Semi-supervised Relation Extraction with Large-scale Word Clustering" pptx

... units with predefined class labels. Given an input resume T which is a sequence of words w 1 ,w 2 ,…,w k , the result of general information extraction is a sequence of blocks in which some words ... problems may occur when estimating either P(T|L) with unknown word w i or P(L) with unknown events. Bikel et al. (1999) mapped all unknown words to one token _UNK_ and then use...

Ngày tải lên: 20/02/2014, 15:20

... tokens with prefix “r:”. We use tokens to encourage overlap between different entities. For the word features, we extract all the words be- tween the two arguments, removing stopwords and the words with ... letters. Words with capital letters are usually named entities, and they do not tend to indicate relations. We also extract neigh- boring words of source and destination arguments....

Ngày tải lên: 23/03/2014, 14:20

... extract relational infor- mation of two kinds: ranked relational terms as keywords and surface patterns. Here surface pat- terns are generated with support of ranked rela- tional terms. 4.3.1 Relational ... classified words into Content Words (CWs) and Functional Words (FWs). From each snippet sentence, the entitled concept, related con- cept, or the keyword k cp is considered to be a Con- t...

Ngày tải lên: 23/03/2014, 16:21

... feature templates as above, with the following additional ones to capture contextual information: CWORDS: Indicator function for the previous and next k (= 1, 2, 3) words. CPOS: Indicator function ... local struc- ture prototypes (e.g., in citation extraction, the word pp. always corresponds to the PAGES field, while proceedings is always associated with BOOKTITLE or JOURNAL), or as cer...

Ngày tải lên: 30/03/2014, 17:20

... detection has been mainly associated with the ICLE, published in 2002. Koppel et al (2005) first constructed such a system with a feature set consisting of function words, POS bi-grams, and character ... dis- tributions of unbounded size with a dynamic param- eter set that grows with the size of the training data. One method of TSG induction is to represent a prob- abilistic TSG with...

Ngày tải lên: 07/03/2014, 18:20

... verb relations, into event-based summarization. Different from other thesaurus like WordNet, VerbOcean provides five types of semantic verb relations at finer level. This just fits in with ... derived from documents and generally associated with single events, but they neglect the relations among events. However, events are commonly related with one another especially when the do...

Ngày tải lên: 23/03/2014, 18:20

... allosteric DNAzyme with dual RNA-cleaving and DNA-cleaving activities is very challenging. We used an in vitro-selected pistol-like (PL) DNAzyme as a DNA scaffold for designing a DNAzyme with dual catalytic ... As compared with PL DNAzyme, DRc DNA- zyme function seemed to have decreased sensitivity to the reaction temperature. DRc DNAzyme exhibited a broad functional range for temperatu...

Ngày tải lên: 16/02/2014, 15:20

... cisplatin DNA treatment correlated with the occurrence of sequence motifs forming stable bifunc- tional adducts with the drug (namely, GG and AG doublets) within the target sites. Binding of ... protein with cer- tain response elements can be controlled by changes in DNA topology inducing formation of non-B DNA structures within the binding sites [11,12]. Interactions of p53 with DNA...

Ngày tải lên: 19/02/2014, 05:20

... Biochem. 271, 1250–1257 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04025.x relation with xenobiotic metabolism or with metabolism of endogenous compounds [24,25]. Models for CYP51 were also ... the RDDT R strain were measured with all recombinant pro- teins, namely testosterone hydroxylation and DDT meta- bolism. Testosterone was hydroxylated at the 2b position with little quan...

Ngày tải lên: 19/02/2014, 12:20

... of translation edit rate with targeted human annotation. In AMTA. Valentin I. Spitkovsky, Hiyan Alshawi, and Daniel Juraf- sky. 2011. Lateen em: Unsupervised training with multiple objectives, ... different metrics M k (h) for evaluating the quality of h. Without loss of gen- erality, we assume metric scores are bounded be- tween 0 and 1, with 1 being perfect. Each hypoth- esis h can be...

Ngày tải lên: 19/02/2014, 19:20