Skip to main navigation Skip to search Skip to main content

Linguistic-first approach to learning Python for natural language generation: Problem breakdown to pseudocode

  • John Blake*
  • , Kazuma Tamura
  • , Evgeny Pyshkin
  • *Corresponding author for this work
  • Center for Language Research (CLR)

Research output: Chapter in Book/Published conference outputConference publication

Abstract

Participants in an elective course on natural language generation developed their ability to program in Python through a linguistic-first, problem-based approach. The primary aim of the course was to create a text generation program, but in doing so, students achieved the secondary aim of increasing their mastery of Python. The course started with a thorough linguistic analysis of the genre of the target language, using both top-down and bottom-up approaches. This served as the basis for the development of a set of guiding principles. These principles were then used to develop pseudocode, which, in turn, served as the foundation for the initial draft of the program. Lessons learned include the importance of aligning aims and assessment criteria, and providing learners with space to struggle.

Original languageEnglish
Title of host publicationAIP Conference Proceedings
EditorsDebopriyo Roy, George Fragulis
Volume2909 (1)
DOIs
Publication statusPublished - 28 Nov 2023
Event5th International Conference on ICT Integration in Technical Education, ETLTC 2023 in collaboration with the 2nd International Conference on Entertainment Technology and Management, ICETM 2023 - Aizuwakamatsu, Japan
Duration: 24 Jan 202327 Jan 2023

Publication series

NameAIP Conference Proceedings
Number1
Volume2909
ISSN (Print)0094-243X
ISSN (Electronic)1551-7616

Conference

Conference5th International Conference on ICT Integration in Technical Education, ETLTC 2023 in collaboration with the 2nd International Conference on Entertainment Technology and Management, ICETM 2023
Country/TerritoryJapan
CityAizuwakamatsu
Period24/01/2327/01/23

Fingerprint

Dive into the research topics of 'Linguistic-first approach to learning Python for natural language generation: Problem breakdown to pseudocode'. Together they form a unique fingerprint.

Cite this