Embedding-based system for the Text part of CALL v3 shared task
This work addresses the challenge of automated language assessment for educational applications, showing incremental improvement by eliminating the need for predefined correct answers.
The paper tackled the problem of scoring text responses in the CALL v3 shared task without using a reference grammar file, achieving top results on the text subset by employing text embeddings like NNLM and BERT.
This paper presents a scoring system that has shown the top result on the text subset of CALL v3 shared task. The presented system is based on text embeddings, namely NNLM~\cite{nnlm} and BERT~\cite{Bert}. The distinguishing feature of the given approach is that it does not rely on the reference grammar file for scoring. The model is compared against approaches that use the grammar file and proves the possibility to achieve similar and even higher results without a predefined set of correct answers. The paper describes the model itself and the data preparation process that played a crucial role in the model training.