Question Answering Systems Approaches and Challenges

Abstract

Question answering (QA) systems permit the user to ask a question using natural language, and the system provides a concise and correct answer. QA systems can be implemented for different types of datasets, structured or unstructured. In this paper, some of the recent studies will be reviewed and the limitations will be discussed. Consequently, the current issues are analyzed with the proposed solutions.

Deep Learning Based Question Answering System (Survey)

“…The need for a lot of labelled data to train the algorithms is one of the primary restrictions. This can be difficult in fields where getting annotated data is difficult or expensive [58]. A further drawback is the inability of deep learning models to be interpreted, resulting in it challenging to comprehend how a model generates its results.…”

“…Natural Language Processing (NLP) approaches have several limitations in Question Answering (QA) systems. The intricacy of feature design is one of the key drawbacks, however deep learning techniques can get around this [104]. The accurateness of the system may also be hampered by the level of detail of the supplied data [105].…”

“…Although neural network-based approaches have been extensively used in Question Answering (QA) systems, they do have significant drawbacks. The intricacy of obtaining features is one of the key drawbacks, however deep learning techniques can get over this [108]. Massive amounts of labelled data are a further constraint that certain applications may find difficult to meet [109].…”

“…Likewise these systems' productivity is still beneath what people would anticipate, and unresolved QA systems' responses are frequently general and call for additional investigation [110]. Deep learning techniques, such as determining feature intricacy, are capable of getting around some of these constraints [108]. Although neural networkbased approaches have helped QA systems operate more effectively in general, there are still several issues that must be resolved for them to operate at their best.…”

Enabling a Question-Answering System for COVID Using a Hybrid Approach Based on Wikipedia and Q/A Pairs

“…Question-answering is a common NLP task. According to Alqifari [1], p. 1, a QA is a "type of system in which a user can ask a question using natural language, and the system provides a concise and correct answer." Therefore, the main problem of QA is, given a set of information resources r and a user's query q, we need to find the best answer a for q.…”