A Simple Surface Realizer for Filipino

18885

Authors

nan

Abstract

. Surface realizers are used at the final phase of natural language text generation to convert abstract or symbolic representations of information to linguistic forms in a particular human language. Rules of grammar for the target language are applied to produce text that is syntactically, morphologically and orthographically correct. In this paper, we present the development of FilSuRe, a simple surface realizer for Filipino. We also present how the Booklat system was produced by revising the realization phase of a prototype automatic story generator system, Picture Books, to use the API library of FilSuRe in order to generate children’s stories in Filipino. Keywords: Surface realization, text generation, story generation, Filipino language. ∗ Copyright 2011 by Ethel Ong, Stephanie Abella, Lawrence Santos, and Dennis Tiu. 1

Keywords

nan

Full Content

A Simple Surface Realizer for Filipino ∗ Ethel Ong, Stephanie Abella, Lawrence Santos, and Dennis Tiu Center for Language Technologies, College of Computer Studies De La Salle University, Manila, Philippines 1004 ethel.ong@delasalle.ph, stephanie.abella@yahoo.com, lawrencesantos@ymail.com, den_tiu@yahoo.com Abstract. Surface realizers are used at the final phase of natural language text generation to convert abstract or symbolic representations of information to linguistic forms in a particular human language. Rules of grammar for the target language are applied to produce text that is syntactically, morphologically and orthographically correct. In this paper, we present the development of FilSuRe, a simple surface realizer for Filipino. We also present how the Booklat system was produced by revising the realization phase of a prototype automatic story generator system, Picture Books, to use the API library of FilSuRe in order to generate children’s stories in Filipino. Keywords: Surface realization, text generation, story generation, Filipino language. ∗ Copyright 2011 by Ethel Ong, Stephanie Abella, Lawrence Santos, and Dennis Tiu. 1 Introduction Text generation systems are computer applications that use theories of artificial intelligence and computational linguistics to automatically produce documents, reports, explanations, help messages, stories, and other kinds of texts. Such systems typically follow a basic three-stage pipeline presented by Dale and Reiter (2000) comprising of content determination, sentence planning, and surface or linguistic realization. Surface or linguistic realization is a process whereby abstract or symbolic representations of information are transformed into readable text in a target human language. A number of systems that generate text rely on an external surface realizer during their realization phase to form syntactically and morphologically correct sentence structures. SimpleNLG (Venour and Reiter, 2007) is a popular choice as its library of Java functions provides an interface for applications to build phrases and simple sentences, and to perform inflectional morphological operations and orthography to generate grammatically correct and formatted English sentences. It is a grammar-based realizer engine that accepts canned and non-canned input representations and produces an output string in a deterministic fashion (Gatt and Reiter, 2009). Various systems have utilized simpleNLG in the latter phase of their generation process. The automatic story generators Picture Books (Hong et al., 2009; Ang et al., 2011) used simpleNLG to transform text specifications in the abstract story tree into story text that can be read by young children. The learning tool for spelling and vocabulary, Pun World (Aban et al., 2010), as well as the learning tool for Filipino heritage learners, SalinLahi (Cheng et al., 2009) utilized simpleNLG to perform formatting on the generated feedback before the surface text is displayed in the user interface. The text generator system, Vigan (Chen et al., 2008), used simpleNLG to transform internal descriptions of museum objects into surface text. Although the functions provided by simpleNLG to produce a sentence in the target language is flexible enough to accept phrases (i.e., noun phrase, verb phrase, prepositional phrase), the structure of a Filipino sentence necessitates the need for a Filipino text realizer that can support the nuances of Filipino grammar rules, specifically those …

Content truncated to 500 words

Labeled Text

. Surface realizers are used at the final phase of natural language text generation to convert abstract or symbolic representations of information to linguistic forms in a particular human language. Rules of grammar for the target language are applied to produce text that is syntactically, morphologically and orthographically correct. In this paper, we present the development of FilSuRe, a simple surface realizer for Filipino. We also present how the Booklat system was produced by revising the realization phase of a prototype automatic story generator system, Picture Books, to use the API library of FilSuRe in order to generate children’s stories in Filipino. Keywords: Surface realization, text generation, story generation, Filipino language. ∗ Copyright 2011 by Ethel Ong, Stephanie Abella, Lawrence Santos, and Dennis Tiu. 1

Publication Details

Paper ID:
18885
Year:
2011
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan