A Question Summarization Method based-on Deep Learning in Persian Language
Arash Ghafouri, Iman Barati, Mohammad Hossein Elahimanesh, Hamid Hasanpour · 2023
To properly answer a question, you must first understand the question. Users typically send more explanations than they need to answer in order to express their question in natural language, which increases the complexity of the question and provides useless side information, which leads to the production of false and unrelated answers and makes it difficult to answer the questions. In this paper, we propose a method for summarizing questions in Persian based on the multilingual pre-trained text-to-text transformer model. To begin, we collected a number of question and question summary pairs from websites for answering religious questions, and after fine-tuning the mT5 model with this dataset, we applied the evaluation criteria of Rouge-1, Rouge-2, and Rouge-L. We examined it using the F-measure and obtained satisfactory results.