S. Pan and K. McKeown. Learning intonation rules for Concept-to-Speech generation. In Proceedings of COLING/ACL'98, Montreal, Canada, 1998.  Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Training a Sentence Planner for Spoken Dialogue Using Boosting - Walker, Rambow, Rogati (2002) (Correct)
.... a corpus of nominal expressions [35, 6] Classifiers have also been trained on corpora labelled for TBI accents to predict the appropriate prosody to output; these prosodic predictors have used various types of input features such as rhetorical structure, semantic features and syntactic features [15, 34]. In addition, some work on stochastic generation has been done within a template based generation paradigm. Walker et. al use reinforcement learning to learn to select among a set of templates to achieve the communicative goals of summarizing or reading a set of email messages [46, 45] Oh and ....
Shimei Pan and Kathleen McKeown. Learning Intonation Rules for Concept to Speech Generation. In COLING-ACL98, pages 1003-1009, Montreal, Canada, 1998.
Do Speakers Realize the Prosodic Structure They Say They Do? - van Herwijnen, Terken (2001) (2 citations) (Correct)
....is made that the annotators are able to accomplish this task well and that the prosodic structure indicated by the annotators corresponds to the prosodic structure of the text when they are actually reading the text aloud. Other studies obtain this HUMAN reference from spoken versions of text [2, 3]. As the latter strategy is far more time consuming we investigate the correspondence between the two types of reference. That is, one reference is obtained from the spoken versions of text, the other reference is obtained from the experts prosodic structures which they predict they would assign ....
Pan, S. and McKeown, K., "Learning Intonation Rules for Concept to Speech Generation", Proceedings of ColingACL '98, Montreal, Canada, 1003-1009, 1998.
The Need for Increased Speech Synthesis Research: Report .. - Sproat, Ostendorf.. (Correct)
....information available to improve speech quality. Other research looks at differences in accuracy for prosody assignment based on part of speech tags, constituency structure, or semantic roles (e.g. agent, patient) when accurately determined by language generation or derived using text analysis [Pan and McKeown, 1998]. While CTS systems to date have exploited information that could conceivably be derived by 21 text analysis techniques though with less accuracy, emerging research explores the potential of information that would be difficult, if not impossible, to derive from text. Semantic and pragmatic ....
Pan, S. and McKeown, K. (1998). Learning intonation rules for concept to speech generation. In Proceedings of COLING/ACL'98, Montreal, Canada.
To appear in Proc. of AAAI,-300 Modeling Prosody.. - Shimei Pan Department (1999) Self-citation (Pan) (Correct)
No context found.
S. Pan and K. McKeown. Learning intonation rules for Concept-to-Speech generation. In Proceedings of COLING/ACL'98, Montreal, Canada, 1998.
Prosody Modeling in Concept-to-Speech Generation.. - McKeown, Pan (2000) (1 citation) Self-citation (Pan Mckeown) (Correct)
....in TTS, such as part of speech (POS) tags or syntactic constituency structure. In these cases, CTS input is more accurate since it was constructed during sentence generation, while TTS input must be approximated from parsing or POS tagging. As a result, we would expect a gain in CTS performance (Pan McKeown 1998). Other information produced in language generation is semantic or pragmatic in nature and often is not available for TTS prosody modeling. CTS prosody modeling might gain the biggest improvements over TTS by modeling this type of information, but it can be di#cult to annotate in speech corpora ....
....participants (obligatory arguments) and one or more circumstances (optional arguments to the verb) Each constituent can have di#erent modifiers, such as classifiers, describers and qualifiers. In earlier work, we experimented with the e#ect that semantic boundary has on various prosodic features (Pan McKeown 1998). The surface realizer uses an English grammar, transforming a lexicalized semantic structure into a syntactic structure, linearizing the structure, and handling morphology and function word generation. The features available after surface realization include syntactic constituent structure, ....
[Article contains additional citation context not shown here]
Pan, S. & McKeown, K. R. 1998 Learning intonation rules for Concept to Speech generation. In Proc. COLING-ACL'98, Montreal, Canada, 1998, pp. 1003--1009.
Learning Intonation Rules for Concept-to-Speech Generation - Pan (1998) (5 citations) Self-citation (Pan) (Correct)
....and syntactic features, discourse features, and deep semantic information. Here, we concentrate on the work we have already done, including building a domain independent prosody model for citation form sentences (isolated sentences) using syntactic and semantic constraints generated by FUF SURGE [ Pan and McKeown, 1998 ] and building a domain specific prosody model for MAGIC using a hierarchical topic structure. Later in Chapter 6, we discuss some of the work we plan to do on building a discourse prosody model and a deep semantic prosody model. All the results presented in this chapter are obtained by using ....
S. Pan and K. McKeown. Learning intonation rules for concept to speech generation. In Proceedings of COLING/ACL'98, Montreal, Canada, 1998.
Modeling Local Context for Pitch Accent Prediction - Pan, Hirschberg (2000) (1 citation) Self-citation (Pan) (Correct)
....more likely to be accented. Other linguistic features, such as inferred given new status (Hirschberg, 1993; Brown, 1983) contrastiveness (Bolinger, 1961) and discourse structure (Nakatani, 1998) have also been examined to explain accent assignment in large speech corpora. In a previous study (Pan and McKeown, 1998; Pan and McKeown, 1999) we investigated how features such as deep syntactic semantic structure and word informativeness correlate with accent placement. In this paper, we focus on how local context in uences accent patterns. More speci cally, we investigate how word collocation in uences ....
S. Pan and K. McKeown. 1998. Learning intonation rules for concept to speech generation. In Proc. of COLING/ACL'98, Montreal, Canada.
Unit Selection and Emotional Speech - Alan Black Language (2003) (2 citations) (Correct)
No context found.
S. Pan, Learning Intonation Rules for Concept-to-Speech Generation, Ph.D. thesis, Columbia University, 1998.
A Survey of Automated Layout Techniques for Information.. - Lok, Feiner (2001) (4 citations) (Correct)
No context found.
S. Pan and K. McKeown. Learning intonation rules for concept to speech generation. In Proc. of the Joint International Conference on Computational Linguistics and Association for Computational Linguistics (COLING-ACL), 1998.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC