Learning Analytics & Machine Learning

Learning analytics represents a field of growing interest amongst researchers. To date, many analytic techniques have been “imported” from related fields such as sociology, statistics, and web science. The appropriations from other fields has produced valuable insight into learning in networks, identifying at-risk learners, and improving analysis of discourse and concept development. A critical next stage for developing the sophistication of LA as a field is to engage with promising research in artificial intelligence, specifically machine learning, fields.

Machine learning (ML) offers learning analytics new opportunities to interact with large scale, often messy, data. ML has gained prominence over the past decade through numerous innovations, including self-driving cars, speech recognition, and advanced web search and productivity tools (such as Google Now). Arthur Samuel’s frequently cited definition of machine learning as giving “computers the ability to learn without being explicitly programmed” is a driving influence in the quest to improve human-computer knowledge development. When applied to learning analytics, machine learning holds promise as a means of not only improved analysis of learner data, but also advancement in identifying at-risk learners, predicting success/failure, improving recommendation systems (both social and content), and moving toward adaptive and personalized learning in even the largest online or in-class learning experiences.

This workshop will introduce students and faculty to machine learning and evaluate opportunities to apply supervised, unsupervised, and semi-supervised learning models to learning analytics. As learning analytics is concerned with a range of challenges, including network analysis, discourse analysis, prediction, adaptivity and personalization, #LAK14ML will explore specific ML solutions to various problems in the learning process and, more broadly, the system of education itself.

This is one of the first workshops attempting to apply ML to LA. In order to accommodate a range of attendee familiarity and expertise, the focus will be on introductory and intermediate concepts and applications.

Follow

Get every new post delivered to your Inbox.