KIS’s Stance Classification Model at the NTCIR-17 QA Lab-PoliInfo-4

We participated in the Stance Classification 2 (SC2) subtask of NTCIR-17 QA Lab-PoliInfo-4 as Team KIS. In this paper, we describe ourstance (agreement or disagreement) classification model for utter-ances of Japanese politicians with domain-adaptive training in thepolitical domain. We additionally trained the Japanese pretrainedLUKE model with a Masked Language Model (MLM) on the Dietminutes dataset. We also preprocessed the model using the head-tail method to truncate utterances longer than the maximum inputlength. We found that these methods were effective, achieved thehighest score of 97.41% in accuracy in the formal run of the sub-task.