Báo cáo khoa học: "Evaluating Response Strategies in a Web-Based Spoken Dialogue Agent" pdf

Tài liệu Báo cáo khoa học: "Co-training for Predicting Emotions with Spoken Dialogue Data" pdf

... instances. We compare the learning curve from Co-training with a baseline of majority class and an upper-bound, in which the classifiers are trained on human-annotated data. Post-hoc analyses ... generate more labeled data automatically. The main focus of this paper is to explore how Co-training can be applied to annotate spoken dialogues. A major challenge to address is in reducing ... more) learners are trained iteratively in tandem. In each iteration, the learners classify more unlabeled data to increase the training data for each other. In theory, the learners must have distinct...

Tài liệu Báo cáo khoa học: "Liars and Saviors in a Sentiment Annotated Corpus of Comments to Political debates" pdf

... opinions? Finally, approaches to opi-nion mining have implicitly assumed that the prob-lem at stake is a balanced classification problem, based on the general assumption that positive and negative ... identifying the most rele-vant challenges in mining opinions targeting media personalities, namely politicians, in comments posted by users to online news articles. We are interested in answering ... mentions are classifiable into syntactic-semantic categories; (vi) the opinionated sentences may be characterized according to their polarity 565and intensity; (vii) each opinionated sentence may...

Tài liệu Báo cáo khoa học: "Identifying Linguistic Structure in a Quantitative Analysis of Dialect Pronunciation" docx

... (Manning and Schütze,1999). Multidimensional scaling is data analysistechnique that provides a spatial display of the datarevealing relationships between the instances in thedata set (Davison, ... multidi-mensional scaling. The analyses showed that resultsobtained using aggregate analysis of word pronunci-ations mostly conform with the traditional phoneticclassiﬁcation of Bulgarian dialects as ... that 3 factors are most im-portant, explaining 35% of the total amount of vari-ance. The main drawback of applying this technique in dialectometry is that it is not directly related to theaggregate...

Tài liệu Báo cáo khoa học: "Mapping Lexical Entries in a Verbs Database to WordNet Senses" doc

... (WSD) in several ways. First, thewords to be disambiguated are entries in a lexicaldatabase, not tokens in a text corpus. Second, wetake an “all-words”rather than a “lexical-sample”approach ... Sim-pleProd, and SimpleWtdSumMajSgl+Aggr: Majority vote of MajSim-pleSgl and MajAggrMajPair+Aggr: Majority vote of MajSim-plePair and MajAggrTable 2 gives recall and precision measures forall variations ... variations at their optimal votethresholds. In the AutoMap+ variation, Grid and Levin+probabilities abstain from voting when their val-ues are zero (a common occurrence, becauseof data sparsity...

Tài liệu Báo cáo khoa học: "Multiple Default Inheritance in a Unification-Based" pdf

... to alternatives at the same conceptual level in the hiersrchy, and in msny cases reflect the traditional ides of 'paradigm'. Equations within a variant set are absolute constraints, ... of class definitions. In equs. tions, unification variables have initial capitals, and negation of constants is indicated by ' '. 'kk' is the string concatenation operator ... information in a more specific class. Each equation in a main set functions as an independent constraint, in a msnner which will be clarified below. Variant equation sets, loosely speaking,...

Tài liệu Báo cáo khoa học: "THEORETICAL/TECHNICAL ISSUES IN NATURAL LANGUAGE ACCESS TO DATABASES" pdf

... appropriate for that database organization. In this way, it would be easy to handle such database organ- izations as that in which the total number of parking lots and/or parking spaces is ... above into a formal query language such as SQL. The relations ZONEF and LUCF are existing database relations, but there is no relation GEOBASE* in the database, giv- ing the subplanning area ... has associated with it (by means of a translation table entry) a relation, which is usually the projection of an existing data base relation. Thus instead of translating a portion of the above...

Tài liệu Báo cáo khoa học: "MACHINE-READABLE COMPONENTS IN A VARIETY OF INFORMATION-SYSTEM APPLICATIONS" potx

...

Tài liệu Báo cáo khoa học: "Domain-transcending mappings in a system for metaphorical reasoning" docx

... SPACEand IDEAS ARE PHYSICAL OBJECTS, con-taining the following relevant mappings:When a person's mind is being viewed as a physical space, an idea's being physically located in the space ... viewneutral mapping adjuncts (VNMAs)1. We regardVNMAs as standard but implicit default aspectsof all view-specific metaphorical mappings. Theyare defaults in that they can, in principle, ... wecall View Neutral Mapping Adjuncts(VNMAs). We give a list of the mainVNMAs that appear to be required, andshow how they can be incorporated into a pre-existing system (ATT-Meta) formetaphorical...

Báo cáo khoa học: "Interpreting Semantic Relations in Noun Compounds via Verb Semantics" pdf

... the open-domain data with smallamounts of training data. Rosario and Marti(2001) achieved about 60% using a neural net-work in a closed domain, Moldovan et al. (2004)achieved 43% using word ... re-vealed that TIME both has an above-average clas-siﬁcation accuracy and is associated with a rela-tively large number of test NCs, while EQUATIVEhas a below-average classiﬁcation accuracy ... irrelevant, i.e. player and coach andcoach and player are equally evidence for an EQUATIVE inter-pretation for player coach (and coach player).accomplish achieve behave conduct ACTact conduct...

Báo cáo khoa học: "Unsupervised Coreference Resolution in a Nonparametric Bayesian Model" potx

... Notethat we take a Bayesian approach in which all pa-rameters are integrated out (or sampled). The infer-ence task is thus primarily a search problem over theindex labels Z.849 (a) (b)(c)The ... is a natural penalty foreach cluster which is actually used.With Z observed during sampling, we can inte-grate out β and calculate P (Zi,j|Z−i,j) analytically,using the Chinese restaurant ... particularly inefﬁcient, so instead wemaintain soft counts over each of these variables anduse these in place of a hard sampling scheme. Wealso found that correctly accounting for the futureimpact...

Xem thêm