Loading...
Performance evaluation of the SBERT model for automatic short answer grading
Date
2023-12-14
Abstract
Automated Short Answer Grading (ASAG) represents an actively researched domain within the field of automated assessment and blended learning. This study is based on the utilization of a Sentence-Transformer Model for ASAG and the assessment of the model's generalizability across a spectrum of publicly available datasets. These datasets encompass a wide range of subjects, including university-level computer science, Natural Language Processing, and science subjects spanning grades 3 to 8. Remarkably, the SBERT model demonstrated exemplary performance across all the datasets examined. However, it is important to fine tune the model on domain-specific data in order to make it efficient for a particular domain, enabling the model to acquire and comprehend the specialized vocabulary relevant to that domain. Additionally, we delve into key considerations essential for designing a domain-specific, yet generalized model for ASAG.
Supervisor
Description
Publisher
Citation
Human-Centred AI Education & Practice Conference (HCAI-ep '23)
Collections
Files
ULRR Identifiers
Funding code
Funding Information
Sustainable Development Goals
External Link
License
Attribution-NonCommercial-ShareAlike 4.0 International