Treffer: LONG-FORM OPEN-DOMAIN QUESTION-ANSWERING SYSTEM ARCHITECTURE.

Question Answering is one of the challenging points of research in natural language processing recently. The problem of automating the answering process for the user's queries became required. So, there were several papers suggested different system architectures for building a question answering systems. In this research paper, we suggest our own system architecture taking into consideration that the input of the system architecture is only the asked question. The suggested system architecture is a long-form open domain question answering that contains mainly two layers. The natural language processing layer which holds the data module and the computing module. This layer is responsible for many operations like pre-processing, preparing, storing the data along with taking the user's question then providing the suitable answer. The dataset of the proposed system has to be documents annotated with questions and answers extracted from these documents. Also, it has to be in SQUAD format. The computing module is a retriever-reader based deep learning model. This model achieves scores: 67% Recall@100 using dense passage retriever model and 67.7% F1 score for reader model. the Interface layer is the second layer which includes the APIs module and the user-interface module. Finally, we will discuss a real time case study for the system. [ABSTRACT FROM AUTHOR]

Copyright of International Journal of Intelligent Computing & Information Sciences is the property of Ain Shams University, Faculty of Computer & Information Sciences and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)