Báo cáo khoa học: "Neural Network Probability Estimation for Broad Coverage Parsing" doc

... Neural Network Probability Estimation for Broad Coverage ParsingJames HendersonDepartement d'InformatiqueUniversite de GeneveJames.Henderson@cui.unige.chAbstractWe present a neural -network- based ... (becoming [S-VP ]). These transformsare undone before any evaluation is performed on the outputtrees. We do not believe these transforms have a major impacton performance, but we have not currently ... method for automati-cally inducing a finite set of features for represent-ing the derivation history. The method is a formof multi-layered artificial neural network calledSimple Synchrony Networks...

Báo cáo khoa học: "Neural Network Recognition of Spelling Errors" docx

... input/output form fixed-length arrays of numbers. A common view in the neural network community has been that different forms of input/output representations have negligible effect on ANN performance. ... spell-checking and name searching. In order for ANNs to recognize variant and homographic forms of words, however, words must be transformed to a form that is meaningful to ANNs. The interface ... on ANN performance. Additionally, different types of word representation are shown to perform better on different types of error. Introduction ANNs are a promising technology for NLP, since...

Tài liệu Báo cáo khoa học: "Performance Conﬁdence Estimation for Automatic Summarization" ppt

... examples for good and badperformance. We also extend the analysis to sin-gle document summarization, for which predict-ing system performance turns out to be much moreaccurate than for multi-document ... systemperformance for a given input is in fact relevant notonly for summarization, but in general for all ap-plications aimed at facilitating information access.In question answering for example, ... summarization than inmulti-document. For example, in the normalizeddata KL divergence has correlation of 0.28 for multi-document summarization but 0.40 for sin-gle document. The number of signature...

Báo cáo khoa học: "Joint and conditional estimation of tagging and parsing models∗" docx

... outperformed the conditionalmodels by quite large amounts. This suggeststhat it may be worthwhile investigating meth-ods for maximum (joint) likelihood estimation for model classes for which ... maximum likelihood estimation and max-imum conditional likelihood estimation for threedifferent kinds of models: PCFG parsers, HMMtaggers and shift-reduce parsers. The results for the PCFG parsers ... (Johnson (1998) reports similiarperformance degradation for other binarizationtransformations). Both the joint and the condi-tional shift-reduce parsers performed much worsethan the PCFG. This...

Báo cáo khoa học: "Bayesian Network, a model for NLP?" ppt

... traditional systems. This ap-proach stands on the formalism, still little ex-ploited for NLP, of Bayesian Network (BN). Asa probabilistic formalism, it offers a great expres-sion capacity ... HCrules method are marked in grey, white is for the LC rules method and black for the Clement’smethod3. The Pronoun node estimates the de-cision probability for a given it occurence to benon-anaphoric.The ... decision4. T he inference stage exploits therelationships for the propagation of the informa-tion and the BN operates by information reinforce-ment to label a pronoun. We applied all precedentrules...

Báo cáo khoa học: Systems biology: parameter estimation for biochemical models ppt

... exploratory step then hasformed a basis for the parameter space with informationin which directions the objective function decreases. Inthe next step, the pattern move, the informationobtained is ... Kaandorp JA & Blom J (2007)Efﬁcient parameter estimation for spatio-temporalmodels of pattern formation: case study of Drosophilamelanogaster. Bioinformatics 23, 3356–3363.37 Kirkpatrick S, Gelatt ... metaheuristic for parameter estimation in nonlin-ear dynamic biological systems. BMC Bioinformatics 7,483.85 Egea JA, Rodrı´guez-Ferna´ndez M, Banga JR & Martı´R (2007) Scatter search for...

Báo cáo khoa học: "Social Network Extraction from Texts: A Thesis Proposal" doc

... novelway of constructing networks by analyzing text tocapture such interactions or events.Motivation: Typically researchers construct a so-cial network from various forms of electronic in-teraction ... is a vastlyrich network present in text such as the content ofemails, comment threads on online social networks,transcribed phone calls. My work will enrich the111social network that SNA ... network analysis for calculating centrality measure of people does nottake into account interactions that people talk aboutin the content of emails. Such linkages are relevantto the task for...

Tài liệu Báo cáo khoa học: "Generating Impact-Based Summaries for Scientiﬁc Literature" docx

... our methodconsistently outperforms all the baselines. Amongthe baselines, MEAD -Doc is consistently better thanboth LEAD and MEAD -Doc+ Cite. While MEAD- Doc s outperforming LEAD is not surprising, ... surprising that MEAD -Doc also outperformsMEAD -Doc+ Cite as the latter uses both the cita-tion context and the original document. One possi-ble explanation may be that MEAD is not designed for impact ... using thedocument collection.Table 2: Impact-based summary vs. regular summary for the paper “A study of smoothing methods for languagemodels applied to ad hoc information retrieval”.nal document...

Tài liệu Báo cáo khoa học: "Mining Wikipedia Revision Histories for Improving Sentence Compression" docx

... of them for such compres-sions/expansions. We make the simplifying assump-tion that all such edits also retain the core mean-ing of the sentence, and are therefore valid trainingdata for our ... for glareprotection] is effective and will help if youroffice has the fluorescent-light overkill [that’s typical in offices].(4) Prices range from $5,000 [for a microvax2000] to $179,000 [for ... data set for the first time, we hope to be able to learn whichparts of the sentence are more likely to come from4 For instance, here is an expansion seen in the data, wherethe added information...

Tài liệu Báo cáo khoa học: "Exploiting N-best Hypotheses for SMT Self-Enhancement" doc

... phrase-alignment information in each hy-pothesis indicates the word reordering for source sentence. We select the word reordering with the highest posterior probability as the best word reor-dering for ... Chinese-to-English task show performance improvements for all the strategies. Moreover, the combination of the three strategies achieves further improvements and outperforms the baseline by 0.67 ... than those that did not. We call the former “good phrase-pairs” and the later “bad phrase-pairs” for the given source input. Hypothetically, we can reinforce the translation model by appending...

Xem thêm