On Applicability of Neural Language Models for Readability Assessment in Filipino

Michael Ibañez, Lloyd Lois Antonie Reyes, Ranz Sapinit, Mohammed Ahmed Hussien, Joseph Marvin Imperial

Research output: Chapter or section in a book/report/conference proceedingChapter in a published conference proceeding

Abstract

In the field of automatic readability assessment (ARA), the current trend in the research community focuses on the use of large neural language models such as BERT as evidenced from its high performance in other downstream NLP tasks. In this study, we dissect the BERT model and applied it to readability assessment in a low-resource setting using a dataset in the Filipino language. Results show that extracting embeddings separately from various layers of BERT obtain relatively similar performance with models trained using a diverse set of handcrafted features and substantially better than using conventional transfer learning approach.

Original languageEnglish
Title of host publicationArtificial Intelligence in Education. Posters and Late Breaking Results, Workshops and Tutorials, Industry and Innovation Tracks, Practitioners’ and Doctoral Consortium - 23rd International Conference, AIED 2022, Proceedings
EditorsMaria Mercedes Rodrigo, Noburu Matsuda, Alexandra I. Cristea, Vania Dimitrova
PublisherSpringer Science and Business Media Deutschland GmbH
Pages573-576
Number of pages4
ISBN (Print)9783031116469
DOIs
Publication statusPublished - 2022
Event23rd International Conference on Artificial Intelligence in Education, AIED 2022 - Durham, UK United Kingdom
Duration: 27 Jul 202231 Jul 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13356 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference23rd International Conference on Artificial Intelligence in Education, AIED 2022
Country/TerritoryUK United Kingdom
CityDurham
Period27/07/2231/07/22

Keywords

  • BERT
  • Neural language models
  • Readability assessment

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint

Dive into the research topics of 'On Applicability of Neural Language Models for Readability Assessment in Filipino'. Together they form a unique fingerprint.

Cite this