P18-3017.pdf

P18-3017.pdf - Exploring Chunk Based Templates for...

  • No School
  • AA 1
  • 7

This preview shows page 1 - 2 out of 7 pages.

Proceedings of ACL 2018, Student Research Workshop , pages 120–126 Melbourne, Australia, July 15 - 20, 2018. c 2018 Association for Computational Linguistics 120 Exploring Chunk Based Templates for Generating a subset of English Text Nikhilesh Bhatnagar LTRC, IIIT Hyderabad Manish Shrivastava LTRC, IIIT Hyderabad Radhika Mamidi LTRC, IIIT Hyderabad Abstract Natural Language Generation (NLG) is a research task which addresses the auto- matic generation of natural language text representative of an input non-linguistic collection of knowledge. In this paper, we address the task of the generation of grammatical sentences in an isolated con- text given a partial bag-of-words which the generated sentence must contain. We view the task as a search problem (a problem of choice) involving combinations of smaller chunk based templates extracted from a training corpus to construct a complete sentence. To achieve that, we propose a fitness function which we use in conjunc- tion with an evolutionary algorithm as the search procedure to arrive at a potentially grammatical sentence (modeled by the fit- ness score) which satisfies the input con- straints. 1 Introduction One of the reasons why NLG is a challenging problem is because there are many ways in which a given content can be represented. These are rep- resented by the stylistic constraints which address syntactic and pragmatic choices (largely) indepen- dent of the information conveyed. Classically, there are two major subtasks recog- nized in NLG: Strategic Generation and Tactical Generation (Sentence Planning and Surface Real- ization) 1 ( Reiter and Dale , 2000 ). Strategic Gen- eration - “what to say” deals with identifying the relevant information to present to the audience and Tactical Generation - “how to say” addresses the 1 Because we follow a template based approach, there is some overlap between the Content Determination and Ag- gregation steps. problems of linguistic representation of the input concepts. In this work, we address the problem of tactical generation, with a focus on the grammat- icality of the generated sentences. We formulate our task as follows: to generate syntactically cor- rect sentences given a set of constraints such as a bag-of-words, partial ordering, etc. So, for example, given a bag of words such as “man”, “plays”, “football” and length constraints, a sentence like “The man plays football in Octo- ber.” would be acceptable. Our approach involves a corpus derived formu- lation of template based generation. Templates are instances of canned text with a slot-filler structure (“gaps”) which can be filled with the appropriate information thus realizing the sentence. Since they are a manual resource, it is rather expensive and hard to generalize over different types or domains of text.
Image of page 1

Subscribe to view the full document.

Image of page 2

{[ snackBarMessage ]}

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern