An alternate approach towards meaningful lyric generation in Tamil

  • Authors:
  • Ananth Ramakrishnan A;Sobha Lalitha Devi

  • Affiliations:
  • MIT Campus of Anna University, Chennai, India;MIT Campus of Anna University, Chennai, India

  • Venue:
  • CALC '10 Proceedings of the NAACL HLT 2010 Second Workshop on Computational Approaches to Linguistic Creativity
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper presents our on-going work to improve the lyric generation component of the Automatic Lyric Generation system for the Tamil Language. An earlier version of the system used an n-gram based model to generate lyrics that match the given melody. This paper identifies some of the deficiencies in the melody analysis and text generation components of the earlier system and explains the new approach used to tackle those drawbacks. The two central approaches discussed in this paper are: (1) An improved mapping scheme for matching melody with words and (2) Knowledge-based Text Generation algorithm based on an existing Ontology and Tamil Morphology Generator.