Please use this identifier to cite or link to this item:
Title: Towards a Framework for Teaching Artificial Intelligence in Higher Education
Authors: Allen, Becky
Issue Date: 2022
Publisher: Newcastle University
Abstract: Artificial Intelligence and the domain’s sub-disciplines are becoming increasingly prevalent within numerous areas of academia and can now be considered a core area of Computer Science (Shapiro, Fiebrink and Norvig, 2018). As a consequence, the Higher Education (HE) sector are increasing their provision of Machine Learning and Artificial Intelligence courses. However, there is a current lack of research pertaining to the best practice for teaching this complex domain, which relies heavily on both computing and mathematics knowledge. This thesis outlines a review of the current education provision in AI within higher education, assessed through qualitative techniques encompassing both lecturer and student interactions. Through completion of case studies at varying educational institutions, potential barriers to learning were identified including issues with mathematics anxiety and low confidence in technical skills. The thesis introduces MetaLearning, a learning resource created as part of this research to serve as an introductory course for Machine Learning. MetaLearning consists of a framework of topics pertinent to an introductory course. This framework was developed from the findings of the review of the current educational provision which identified key topics for inclusion. MetaLearning also incorporates a number of mitigation strategies to assist learners in overcoming some of the identified barriers. Strategies pertain to improving student’s metacognition and self-efficacy with the overall aim of learners becoming more self regulated, therefore equipping them with the tools to persevere when encountering difficulties such as threshold concepts. A review of MetaLearning, outlining both the student and lecturer view of the efficacy of this resource is included. Finally, an initial framework is outlined for the best practice for teaching AI. This includes issues pertaining to educational background, mathematics anxiety and low self-efficacy. Alongside an initial overview of the potential threshold concepts, guidance to improve student attainment and satisfaction within these courses is also discussed. Although these findings were the outcome of research specific to AI, they have relevance and will generalise to the wider overarching Computer Science domain.
Description: PhD Thesis
Appears in Collections:School of Computing

Files in This Item:
File Description SizeFormat 
AllenB2022.pdfThesis8.46 MBAdobe PDFView/Open
dspacelicence.pdfLicence43.82 kBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.