The ISAAC project
The Interactive Short Answer Assessment Component is a project at the interface of Natural Language Processing and Education designed to provide
- interactive feedback, and
- automated assessment
for short answers to comprehension questions in language learning tasks.
Project summary
Leveraging the possibilities of digitalization in the area of education is one of the main challenges of our time. In the area of language learning, two key technologies play a larger role here:
-
Intelligent Tutoring systems, supporting individualized, sustainable learning with adaptive, immediate feedback and
-
methods for the automatic assessment of complex task types.
In practice, complex open task types with free-text answers can currently only be assessed with significant manual labor. They do however play an important role in determining learner competence in language and content learning. For Intelligent Tutoring systems, research shows that they lead to substantial learning improvements. Individualized technological support of the learner also supports more equality in education, since learners are less dependent on the education level of their parents. On the basis of Ziai (2018)’s research on the automatic Content Assessment of answers to questions, this project aims to make use of the vast potential of Natural Language Processing in a real-life education context by developing two crucial components: an automatic assessment approach for institutional language testing, and an interactive feedback module for individual practice of learners. Concretely, this means that at the end of the funding period, software modules will exist for both components which can directly be integrated into the infrastructure of relevant partner institutions (learning management systems, testing frameworks) and applied in schools.
Project team
Project head
Research assistants
- Zhuge Gao
Former team members
- Anna Karnysheva
- Eric Demattos
- Jannis Spiekermann
Publications
Ramon Ziai and Anna Karnysheva (2021). Leveraging Task Information in Grammatical Error Correction for Short Answer Assessment through Context-based Reranking. In Proceedings of the 10th Workshop on Natural Language Processing for Computer Assisted Language Learning (NLP4CALL 2021), pp. 62–68. LiU Electronic Press / ACL.
Code
All components in ISAAC are developed open-source on GitHub. Questions, statements of interest or collaboration, and contributions are welcome!
Project partners
 |
Supported by
 |
Funding
Funded by the Federal Ministry of Education and Research (BMBF) and the Baden- Württemberg Ministry of Science as part of the Excellence Strategy of the German Federal and State Governments.
 |
 |
 |