Presentation 2021-12-11
Development of trait-based neural automated essay scoring incorporating multidimensional item response theory
Takumi Shibata, Masaki Uto,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In recent years, deep neural network (DNN)-based automated essay scoring (AES) models that can simultaneously predict the overall score and multiple trait-specific scores have been proposed. However, the main problem of conventional models is the lack of explainability. To resolve this problem, this study proposes a new trait-based DNN-AES model with high explainability by integrating multidimensional item response theory. The proposed model succeeded in improving explainability without a significant loss of accuracy compared to a state-of-the-art conventional model.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) essay-type tests / automated essay scoring / deep neural networks / multidimensional item response theory / explainability
Paper # ET2021-33
Date of Issue 2021-12-04 (ET)

Conference Information
Committee ET
Conference Date 2021/12/11(1days)
Place (in Japanese) (See Japanese page)
Place (in English) Online
Topics (in Japanese) (See Japanese page)
Topics (in English) Sessions for Young Researchers (Young Researcher Awards Selection), etc.
Chair Kenji Watanabe(Hiroshimai Univ.)
Vice Chair Yasuhiro Fujihara(Hyogo College of Medicine)
Secretary Yasuhiro Fujihara(Kochi Univ.)
Assistant Sho Yamamoto(Kinki Univ.) / Toru Kano(Tokyo University of Science)

Paper Information
Registration To Technical Committee on Educational Technology
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Development of trait-based neural automated essay scoring incorporating multidimensional item response theory
Sub Title (in English)
Keyword(1) essay-type tests
Keyword(2) automated essay scoring
Keyword(3) deep neural networks
Keyword(4) multidimensional item response theory
Keyword(5) explainability
1st Author's Name Takumi Shibata
1st Author's Affiliation The University of Electro-Communications(UEC)
2nd Author's Name Masaki Uto
2nd Author's Affiliation The University of Electro-Communications(UEC)
Date 2021-12-11
Paper # ET2021-33
Volume (vol) vol.121
Number (no) ET-294
Page pp.pp.23-28(ET),
#Pages 6
Date of Issue 2021-12-04 (ET)