Báo cáo khoa học: "Improving the Use of Pseudo-Words for Evaluating Selectional Preferences" docx

... implement. The Backoff Erk model is the best, using the Baseline for the majority of decisions and backing off to the Erk smoothingmodel when the Baseline cannot answer.Figure 5 (shown on the next ... which of two verbswas the correct predicate for a given noun object.One verb v was the original from the source doc-ument, and the other vwas randomly generated.This was the ﬁrst use of such ... is the concatenation of twowords (e.g. house/car). One word is the orig-inal in a document, and the second is the con-founder. Consider the following example of ap-plying pseudo-words to the...

Tài liệu Báo cáo khoa học: "Improving the Scalability of Semi-Markov Conditional Random Fields for Named Entity Recognition" pdf

... and the former was used as the trainingdata and the latter as the development data. For semi-CRFs, we used amis3 for training the semi-CRF with feature-forest. We used GENIA taggar4 for POS-tagging ... than the system without it (the p-value isless than 1.0 < 10−4). The result of the precedingentity information improves the performance. On the other hand, the system with preceding infor-mation ... uses the infor-mation about the preceding entity and 1.0 ×10−15threshold probability for ﬁltering. We note that the result of our system is similar to those of other sys-5 The result of the...

Tài liệu Báo cáo khoa học: "Exploring the Use of Linguistic Features in Domain and Genre Classification" potx

... vectors: Another lesson of Tab. 3 is that the effect of the com- position of the feature vectors can vary depend- ing both on the task and on the size of the feature vector. The dramatic ... because some classes, such as L, are very small. This problem was not so grave for the LPE experiments because of the ceiling effect and the small size of the complete data set, therefore, ... 90%. Size of Feature Vector: The number of features used did not significantly affect the performance of IBL. For LVQ, both precision and re- call decrease sharply as the number of features...

Báo cáo khoa học: "Improving the Interpretation of Noun Phrases with Cross-linguistic Information" doc

... the contribution the features exempliﬁed in one baselineand six versions of the SVM model. The baseline isdeﬁned only for the English part of the NP featureset and measures the the contribution of the ... EXPERIENCER, THEME, BENEFICIARY.Out of these instances, 74.81% use the preposition of. In CLUVI, 11.71% of the examples were ver-bal, from which the preposition of has a coverage of 82.20%. The many-to-many ... of the prepo-sitions (especially of/ de) to the semantic classes addsto the complexity of the interpretation task. Thus, for the interpretation of these constructions a systemmust rely on the...

Báo cáo khoa học: "Improving the Accuracy of Subcategorizations Acquired from Corpora" pdf

... lexicon of the tar-get grammar,4and make use of the existing sets of 4When the lexicon is less accurate, I can determine the number of clusters using other algorithms (Hamerly, 2003).SCFs for the ... vectors from the trainingSCFs and the acquired SCFs for the words in the testing SCFs. The number of the resulting dataobjects was 8,679 for XTAG and 1,241 for ERG. The number of initial centroids7extracted ... expectations.8Precision=Correct SCFs for the words in the resulting SCFsAll SCFs for the words in the resulting SCFsRecall =Correct SCFs for the words in the resulting SCFsAll SCFs for the words in the test SCFs...

Báo cáo khoa học: "Improving the Performance of the Random Walk Model for Answering Complex Questions" pptx

... cases where the query is composed of two or more sentences, wecompute the similarity between the document sen-tence (s) and each of the query-sentences (qi) thenwe take the average of the scores.3 ... butat the same time makes it not well suited for the se-mantic trees (ST) deﬁned in Section 3. For instance,although the two STs of Figure 1 share most of the subtrees rooted in the ST node, the ... represents the occurrencefrequency (TF*IDF) of a word. However, the majorlimitation of the TF*IDF approach is that it only re-tains the frequency of the words and does not takeinto account the sequence,...

Báo cáo khoa học: "On the use of Comparable Corpora to improve SMT performance" ppt

... on the English part of the bitexts and the Gigaword corpus of about3.2 billion words. Therefore, it is likely that the target language model includes at least some of the translations of the ... amounts of parallel textsto translate the source side of the non-parallel corpus. The target side texts areused, along with other corpora, in the lan-guage model of this SMT system. Wethen use ... Proceedings of the 12th Conference of the European Chapter of the ACL, pages 16–23,Athens, Greece, 30 March – 3 April 2009.c2009 Association for Computational LinguisticsOn the use of Comparable...

Báo cáo khoa học: "Comparing the Accuracy of CCG and Penn Treebank Parsers" docx

... is the number of sentencesin the sample, and the % column gives the samplesize as a percentage of the whole section.We compare the CCG parser to the Berkeleyparser using the accurate mode of ... closer to the PTB than CCGbank, or due to theirconversion method. We leave the application of their methodto the CCG parser for future work.to use the complete test set. However, there are ... rules and for the rules dealing withpunctuation, which allowed most of the 2853 ruleinstances to be covered. Considerable time and ef-fort was invested in the creation of these schemas. The oracle...

Báo cáo khoa học: "ON THE INDEPENDENCE OF DISCOURSE STRUCTURE AND SEMANTIC DOMAIN" docx

... layouts, there is a minority strategy, used by 4% of the speakers (3 out of 72 cases of the data of Linde (1974)) describing the layout in the form of a map. The speaker first describes the outside ... describe the layout of their apartment. The vast majority of speakers used a "tour strategy," which takes the hearer on an imaginary tour of the apartment, building up the description of ... breadth first traversal with the root being the outside shape, the internal divisions the next layer of nodes, and the names of these divisions the terminal nodes. Because there are so few example,...

Tài liệu Báo cáo khoa học: Seeking the determinants of the elusive functions of Sco proteins pptx

... function of the CXXXCmotif of human Sco proteins could therefore be impli-cated not only in the maturation of the CuAsite of Cox2 but also in the maintenance of cellular copperhomeostasis. The ... assembly of cbb3oxidase, but rather isrequired for the maturation of the CuA-containingCOX which is predominant for aerobic growth, thusleaving open the question as to the identity of the oxi-dase ... binding and concomitantformation of the compact Cu(I)-Sco1 structure. The formation of the stable compact Cu(I)-Sco1 state couldthus constitute the important driving force of the cop-per transfer...

Xem thêm