What is it about?
Question Answering System (QAS) is a domain of Natural Language Processing which is built on the foundation of Machine Learning. Do machines behave at par with humans while answering the questions? In our proposed system, we have shown how the machine can think whether the question provided by the user is appropriate and accordingly answer that question.
Featured Image
Why is it important?
The performance of the state-of-the-art QAS models depends upon how the model answers the answerable questions. Currently, the models are being evaluated by making them answer for both answerable and unanswerable questions. The proposed system has improved the reasoning power of the QAS by introducing a method called “Question Similarity Mechanism,” which identifies and filters the unanswerable and irrelevant questions from being posed to the QAS. Also, this whole system can be used for real-time applications, wherein for a given text, a list of questions with appropriate answers is produced automatically.
Perspectives
Pre-trained models are the leader boards in NLP-related tasks. Even though these models generate good results, sometimes they generate factually incorrect results. The state-of-the-art models in the QAS are pre-trained. In this work, we focus on one of the major limitations of the QAS models and develop a mechanism that effectively filters the unanswerable and irrelevant questions from being posed to QAS. We hope that this research will provide some insights to the researchers, who are working in the domain of Question Answering System.
Dr. Sanjay Singh
Manipal Institute of Technology, Manipal
Read the Original
This page is a summary of: Automatic question-answer pairs generation and question similarity mechanism in question answering system, Applied Intelligence, April 2021, Springer Science + Business Media,
DOI: 10.1007/s10489-021-02348-9.
You can read the full text:
Contributors
The following have contributed to this page
