Please use this identifier to cite or link to this item: https://www.um.edu.mt/library/oar/handle/123456789/91981
Title: A.I. assisted pedagogical chatbot
Authors: Attard, Andrew Emanuel (2021)
Keywords: Artificial intelligence
Intelligent tutoring systems
Computer-assisted instruction
Natural language generation (Computer science)
Issue Date: 2021
Citation: Attard, A. E. (2021). A.I. assisted pedagogical chatbot (Bachelor’s dissertation).
Abstract: When the traditional classroom setting is examined, an observer can immediately observe a paradigm where an educator is tasked with transferring knowledge to multiple students at the same time. This paradigm is fundamentally flawed due to the number of students who need constant attention and educational monitoring throughout their studies while there is only one teacher in a classroom at a given time. This work is focused around on Mathematics where students find more difficulty with the topic as it requires more hands-on experience and thus there is a higher demand for individual attention. This project aims to solve this issue by enabling educators and their students to make use of a chatbot that has been generated based on the educator’s own words. The chatbot is created by obtaining basic Mathematical explanations from the educator and then generating multiple explanations that are similar in content and wording. When prompted the chatbot will then supply the appropriate explanation based on the question that it is matched to. The data necessary to generate a chatbot is obtained by extracting possible questions from the explanations given by the educator, the purpose of these questions is to simulate user input. The responses of the chatbot (i.e. the explanations) are a collection of the originally supplied explanations as well as a set of the generated explanations. When generated, the explanations are selected based on their quality which is evaluated by a Multinomial Naive-Bayes classifier, furthermore, once the explanations have been selected, versions of the explanations with different numbers are generated to provide chatbot with a wide variety of responses. The results of this study were evaluated via an online survey where several voluntary participants were recruited online and given the chatbot to test and then fill in the following survey. The survey indicates that 73% of the users enjoyed making use of the chatbot, and the same percentage of respondents also expressed a desire to use the chatbot again in the future.
Description: B.Sc. IT (Hons)(Melit.)
URI: https://www.um.edu.mt/library/oar/handle/123456789/91981
Appears in Collections:Dissertations - FacICT - 2021
Dissertations - FacICTAI - 2021

Files in This Item:
File Description SizeFormat 
21BITAI003.pdf
  Restricted Access
2 MBAdobe PDFView/Open Request a copy


Items in OAR@UM are protected by copyright, with all rights reserved, unless otherwise indicated.