We offer a preliminary study of how knowledge may be represented differently in different languages. In particular, we account for the contrast between English and Chinese when identical target domain knowledge is rep...
详细信息
We offer a preliminary study of how knowledge may be represented differently in different languages. In particular, we account for the contrast between English and Chinese when identical target domain knowledge is represented with two different, yet related, source domains in each language. We incorporate corpora analysis in English and Chinese with SUMO to delimit the source domains identified using the conceptual mapping model (Ahrens. 2002). In particular, we investigate economy metaphors with the source domains of aeroplane and moving vehicle. These two source domains are found in Chinese and English respectively. Hence, we ask the question whether these source domains should be conflated under a general source domain such as TransportationDevice or should they form independent source domains. Our study addresses this issue by using the SUMO ontology as well as the mapping principles found in the corpora analysis. Our studies contribute to automatizing the source-target domain mappings in conceptual metaphors.
＜正＞A framework is proposed for enterprise automated call routing system development and large scalable naturallanguage call routing application deployment based on IBM’s speech recognition and NLU application eng...
详细信息
＜正＞A framework is proposed for enterprise automated call routing system development and large scalable naturallanguage call routing application deployment based on IBM’s speech recognition and NLU application engagement practices in recently years. To facilitate employing different call classification algorithms in an easy integration manner,this framework architecture provides a plug ＆ play environment for evaluating promising call routing algorithms and a systematic approach to carry out a large scalable enterprise application deployment. The paradigm in this paper illustrates the complementary effort to develop an automatic call routing application for enterprise call centers and covers from call classification algorithm investigation to application programming model. Experimental results on a live data testing set collected from an enterprise call center shows that the performance of the call classification algorithm implemented in this framework is outstanding.
＜正＞In this paper,two methods for Intonational Phrase（INP） prediction in Chinese TTS（Text To Speech） system are proposed,both of which exploit the information of Base Phrase（BP）.Method 1is called BP-based INP ...
详细信息
ISBN:
(纸本)0780379020
＜正＞In this paper,two methods for Intonational Phrase（INP） prediction in Chinese TTS（Text To Speech） system are proposed,both of which exploit the information of Base Phrase（BP）.Method 1is called BP-based INP boundary prediction,in which,BPs in a sentence are recognized at first,then INP boundaries are predicted on the basis of words and recognized BPs Method 2 is called BP-filtering INP boundary prediction,in which, recognized BPs are used to filter out some impossible *** Tree（DT） was used as learning method for both BP recognition and INP boundary piediction. Comparing to the conventional method without BP recognition,the two methods proposed here achieved 3.6%and 5.6%reduction in the measure of unacceptability *** analysis shows that, by improving the performance of BP recognition,there is maximally 36.9%reduction in unacceptability.
Ontology has become increasingly important in lexical semantics and languageprocessing. Researchers also aim at constructing a universal linking network to represent human *** taking verbs of communication as an exam...
详细信息
ISBN:
(纸本)0780379020
Ontology has become increasingly important in lexical semantics and languageprocessing. Researchers also aim at constructing a universal linking network to represent human *** taking verbs of communication as an example,in this paper we discuss problems connecting a verbal domain ontology and an upper level *** taxonomies of the communication domain in the lexical *** and HowNet,are considered as the domain internal ontology and are attempted to map to the Suggested Upper Merged Ontology(SUMO).We propose that there be an interface bridging upper and domain internal ontologies,and that the interface can be analyzed under the Module-Attribute Representation of Verbal Semantics(MARVS) framework in the hope of facilitating the ontology linking network.
＜正＞With Information exploding on the Internet,existing search engines encounter difficulty in accurate document *** content-based search engines are in need for helping us find useful information accurately and ***...
详细信息
ISBN:
(纸本)0780379020
＜正＞With Information exploding on the Internet,existing search engines encounter difficulty in accurate document *** content-based search engines are in need for helping us find useful information accurately and *** on our self-established conceptual network platform ConNet, a novel information retrieval framework is discussed in this paper,which is being used to support semantic matching and concept based searching for multiple domains and multiple languages across widely distributed Internet *** developed ConNet platform under the Conceptual Network Model（CNM）,and mid-stage experiments have shown some encouraging results to illustrate our idea.
We have developed in the last 4 years several platforms allowing to handle various aspects of spontaneous,general-purpose spoken bilingual dialogues on the web:human interpretation,data collection,and integration of m...
详细信息
We have developed in the last 4 years several platforms allowing to handle various aspects of spontaneous,general-purpose spoken bilingual dialogues on the web:human interpretation,data collection,and integration of machine aids including server-based speech translation based on commercial products and user *** data should be available on the web in autumn *** platforms are being integrated into the single ERIMM platform,which should then be extended to be usable for e-learning in interpretation.
In the domain of text categorization(TC),the TF (term frequency)*IDF(inverse document frequency) weighting algorithm and TF*IWF*IWF weighting algorithm are widely ***,the two algorithms are too biased by the term freq...
详细信息
ISBN:
(纸本)0780379020
In the domain of text categorization(TC),the TF (term frequency)*IDF(inverse document frequency) weighting algorithm and TF*IWF*IWF weighting algorithm are widely ***,the two algorithms are too biased by the term frequency and neglect the unbalance between *** this paper, we propose a new weighting algorithm,which is named as TF(term frequency) *IWF(inverse word frequency) *IWF(inverse word frequency)*VE (variance and expectation).The new algorithm improves the TF*IWF*IWF weighting algorithm in both TF and *** paper compares the new algorithm with TF*IWF*IWF algorithm respectively in theory and *** the preliminary experiment,we find that the F1 -Measure has been improved for 11.78%.
The paper describes a rule-learning approach towards Chinese prosodic phrase prediction for TTS systems. 3167 sentences with two-level prosodic phrase labeling information was prepared for analysis. Candidate features...
详细信息
ISBN:
(纸本)0780379020
The paper describes a rule-learning approach towards Chinese prosodic phrase prediction for TTS systems. 3167 sentences with two-level prosodic phrase labeling information was prepared for analysis. Candidate features related to prosodic phrasing were extracted from the corpus to establish an example *** on this,a series of comparative experiments is conducted to collect the most effective features from the *** typical rule learning algorithms(C4.5 and TBL) were applied on the example database to induce prediction *** compare the results with others,the general evaluation parameters were introduced in the paper. With these parameters,the methods were compared to RNN and bigram based *** shown that the rule-learning approach introduced in the paper can achieve better prediction accuracy than the non-rule based methods and yet retain the advantage of the simplicity and understandability
作者：
Kenzo IwamaZA
Corp.3-17-18-106 Makinosaka Hirakata Osaka Japan
This article describes a virtual robot that acquires meanings of words,and discusses that the robot shows an introspective conscious *** article first argues what outputs of the robot demonstrate that it has learned m...
详细信息
This article describes a virtual robot that acquires meanings of words,and discusses that the robot shows an introspective conscious *** article first argues what outputs of the robot demonstrate that it has learned meanings of words such as "what is this".Then the article explains functions of the robot that enable the robot to make outputs as if it has learned meanings of *** functions include 1) forming of linked lists(associations) among observed features,robotic actions,and words,2) generalizing the associations to form generalized associations,and 3) applying an association and a generalized association to a new situation by making the generalized association match words,observed features,and actions in the new situation.
In this paper,we discuss knowledge representation and uncertainty management based on Bayesian Belief Networks(BBNs).Firstly,we carry out an investigation to the identified problems that will occur during applying saf...
详细信息
In this paper,we discuss knowledge representation and uncertainty management based on Bayesian Belief Networks(BBNs).Firstly,we carry out an investigation to the identified problems that will occur during applying safety *** conclude that most of these problems originate from the uncertainty nature in safety *** we provide an in-depth argument to explain why BBNs can deal with these problems,and offer an efficient approach for knowledge representation and uncertainty management for safety assessment. Finally,we present a BBN-based decision-supporting expert system that can be used to assess the implementing quality of overall safety requirement within safety standard IEC *** BBN prototype is based on a systematic identification of failure modes,and modelling of their impacts on the final *** phases of safety-related systems development life cycles are particularly appropriate candidates for this type of analysis.
暂无评论