Bulathwela, Maliththa Sahan Sarojan;
(2023)
Novel Datasets, User Interfaces and Learner Models to Improve Learner Engagement Prediction on Educational Videos.
Doctoral thesis (Ph.D), UCL (University College London).
Preview |
Text
PhD_Thesis_Ready_for_Submission.pdf - Accepted Version Download (14MB) | Preview |
Abstract
With the emergence of Open Education Resources (OERs), educational content creation has rapidly scaled up, making a large collection of new materials made available. Among these, we find educational videos, the most popular modality for transferring knowledge in the technology-enhanced learning paradigm. Rapid creation of learning resources opens up opportunities in facilitating sustainable education, as the potential to personalise and recommend specific materials that align with individual users’ interests, goals, knowledge level, language and stylistic preferences increases. However, the quality and topical coverage of these materials could vary significantly, posing significant challenges in managing this large collection, including the risk of negative user experience and engagement with these materials. The scarcity of support resources such as public datasets is another challenge that slows down the development of tools in this research area. This thesis develops a set of novel tools that improve the recommendation of educational videos. Two novel datasets and an e-learning platform with a novel user interface are developed to support the offline and online testing of recommendation models for educational videos. Furthermore, a set of learner models that accounts for the learner interests, knowledge, novelty and popularity of content is developed through this thesis. The different models are integrated together to propose a novel learner model that accounts for the different factors simultaneously. The user studies conducted on the novel user interface show that the new interface encourages users to explore the topical content more rigorously before making relevance judgements about educational videos. Offline experiments on the newly constructed datasets show that the newly proposed learner models outperform their relevant baselines significantly.
| Type: | Thesis (Doctoral) |
|---|---|
| Qualification: | Ph.D |
| Title: | Novel Datasets, User Interfaces and Learner Models to Improve Learner Engagement Prediction on Educational Videos |
| Open access status: | An open access version is available from UCL Discovery |
| Language: | English |
| Additional information: | Copyright © The Author 2023. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request. |
| Keywords: | AI in Education, Artificial Intelligence, Content Analytics, Educational Data Mining, Educational Recommendation Systems, Learner Modelling, Learning Analytics, Machine Learning, Scalable Quality Assurance |
| UCL classification: | UCL UCL > Provost and Vice Provost Offices > UCL BEAMS UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science > Dept of Computer Science |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10167894 |
Archive Staff Only
 |
View Item |
