EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Multimodal Test Item Parameter Prediction From Text, Images, and Metadata: Fusing Together AI Vision and Language Models

Hotaka Maeda iD, Yikai Ek Lu iD

Educational and Psychological Measurement · 13.07.2026

DOI: 10.1177/00131644261460779 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Educational and Psychological Measurement, online erschienen am 13.07.2026. Verfasst von Hotaka Maeda und 1 weiteren Autor:innen. Beteiligte Einrichtung: Smarter Balanced, Santa Cruz, CA, USA. Die Arbeit nennt 58 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

We propose a flexible multimodal model for predicting all dichotomous and polytomous item parameters from text, images, and metadata by fusing representations from encoder Transformer vision and language models. This deep learning model accommodates heterogeneous item formats, including items with any number of components such as correct and incorrect options, stimuli, and images. Answer key indicators distinguish correct options from distractors, while an attention pooling technique weights the relative importance of these components. Based on the two-parameter logistic model and the generalized partial credit model, we predict all item parameters jointly using a masking strategy to ensure that only relevant parameters contribute to the loss. Item-level and item component-level metadata are also included. We evaluate the approach using 40,965 English language arts and mathematics items for Grades 3–11. A single model accommodated both exam subjects, all 11 item types, and all item parameters, eliminating the need for multiple specialized models. However, results indicate that the full model was unable to leverage all input data. Often, prediction accuracy was unchanged as features were removed. Images were good predictors on their own (item intercept R 2 = . 25 ) but did not consistently contribute unique information when combined with text and metadata. Of the strongest-performing models, the most parsimonious variant achieved R 2 values of .67, .45, .39, .34, .66, .33, and .74 for item intercept, discrimination, difficulty, and four polytomous step threshold parameters, respectively. Findings suggest that current training methods may limit the learnability of complex multimodal deep fusion models. Example Python code from this study is available on Github: https://github.com/hotakamaeda/multimodal_prediction

Schlagwörter

Stichwörter der Autor:innen

DINOv3DeBERTaV3artificial intelligenceitem response theorylarge-scale assessmentnatural language processingquestion difficulty prediction

Autor:innen und Institutionen

  • Smarter Balanced, Santa Cruz, CA, USA.
  • University of California, Santa Cruz, USA.
  • University of Minnesota Twin Cities, Minneapolis, USA.

ORCID-Profile: Hotaka Maeda, Yikai Ek Lu

Zitieren

APA

Maeda, H., & Lu, Y. E. (2026). Multimodal Test Item Parameter Prediction From Text, Images, and Metadata: Fusing Together AI Vision and Language Models. Educational and Psychological Measurement. https://doi.org/10.1177/00131644261460779

Vancouver

Maeda H, Lu YE. Multimodal Test Item Parameter Prediction From Text, Images, and Metadata: Fusing Together AI Vision and Language Models. Educational and Psychological Measurement. 2026. doi:10.1177/00131644261460779

BibTeX
@article{Maeda2026,
  author = {Maeda, Hotaka and Lu, Yikai Ek},
  title = {Multimodal Test Item Parameter Prediction From Text, Images, and Metadata: Fusing Together AI Vision and Language Models},
  journal = {Educational and Psychological Measurement},
  year = {2026},
  doi = {10.1177/00131644261460779}
}
Kostenlos starten →