No abstract

Millions of people have enrolled and enrol (especially in the Covid-19 pandemic world) in MOOCs. However, the retention rate of learners is notoriously low. The majority of the research work on this issue focuses on predicting the dropout rate, but very few use explainable learning patterns as part of this analysis. However, visual representation of learning patterns could provide deeper insights into learners' behaviour across different courses, whilst numerical analyses canand arguably, shouldbe used to confirm the latter. Thus, this paper proposes and compares different granularity visualisations for learning patterns (based on clickstream data) for both course completers and noncompleters. In the large-scale MOOCs we analysed, across various domains, our fine-grained, fish-eye visualisation approach showed that non-completers are more likely to jump forward in their learning sessions, often on a 'catch-up' path, whilst completers exhibit linear behaviour. For coarser, bird-eye granularity visualisation, we observed learners' transition between types of learning activity, obtaining typed transition graphs. The results, backed up by statistical significance analysis and machine learning, provide insights for course instructors to maintain engagement of learners by adapting the course design to not just 'dry' predicted values, but explainable, visually viable paths extracted.

Massive Open Online Courses (MOOCs) have become a popular choice for e-learning thanks to their great flexibility. However, due to large numbers of learners and their diverse backgrounds, it is taxing to offer real-time support. Learners may post their feelings of confusion and struggle in the respective MOOC forums, but with the large volume of posts and high workloads for MOOC instructors, it is unlikely that the instructors can identify all learners requiring intervention. This problem has been studied as a Natural Language Processing (NLP) problem recently, and is known to be challenging, due to the imbalance of the data and the complex nature of the task. In this paper, we explore for the first time Bayesian deep learning on learner-based text posts with two methods: Monte Carlo Dropout and Variational Inference, as a new solution to assessing the need of instructor interventions for a learner's post. We compare models based on our proposed methods with probabilistic modelling to its baseline non-Bayesian models under similar circumstances, for different cases of applying prediction. The results suggest that Bayesian deep learning offers a critical uncertainty measure that is not supplied by traditional neural networks. This adds more explainability, trust and robustness to AI, which is crucial in education-based applications. Additionally, it can achieve similar or better performance compared to non-probabilistic neural networks, as well as grant lower variance.

Massive Open Online Course (MOOC) systems have become prevalent in recent years and draw more attention, a.o., due to the coronavirus pandemic's impact. However, there is a well-known higher chance of dropout from MOOCs than from conventional off-line courses. Researchers have implemented extensive methods to explore the reasons behind learner attrition or lack of interest to apply timely interventions. The recent success of neural networks has revolutionised extensive Learning Analytics (LA) tasks. More recently, the associated deep learning techniques are increasingly deployed to address the dropout prediction problem. This survey gives a timely and succinct overview of deep learning techniques for MOOCs' learning analytics. We mainly analyse the trends of feature processing and the model design in dropout prediction, respectively. Moreover, the recent incremental improvements over existing deep learning techniques and the commonly used public data sets have been presented. Finally, the paper proposes three future research directions in the field: knowledge graphs with learning analytics, comprehensive social network analysis, composite behavioural analysis.

This research focuses on semi-supervised classification tasks, specifically for graph-structured data under datascarce situations. It is known that the performance of conventional supervised graph convolutional models is mediocre at classification tasks, when only a small fraction of the labeled nodes are given. Additionally, most existing graph neural network models often ignore the noise in graph generation and consider all the relations between objects as genuine ground-truth. Hence, the missing edges may not be considered, while other spurious edges are included. Addressing those challenges, we propose a Bayesian Graph Attention model which utilizes a generative model to randomly generate the observed graph. The method infers the joint posterior distribution of node labels and graph structure, by combining the Mixed-Membership Stochastic Block Model with the Graph Attention Model. We adopt a variety of approximation methods to estimate the Bayesian posterior distribution of the missing labels. The proposed method is comprehensively evaluated on three graph-based deep learning benchmark data sets. The experimental results demonstrate a competitive performance of our proposed model BGAT against the current state of the art models when there are few labels available (the highest improvement is 5%), for semi-supervised node classification tasks.