Characterising Video Segments to Support Learning

Abstract

Videos provide opportunities for engagement and independent learning and are widely used in various learning contexts. However, there are challenges with using videos for learning, e.g. long videos can reduce the concentration span, learners may become bored, not everyone can be able to detect the main points in the video, and not all parts in a video will be relevant to the learner. To address these challenges, our research aims to develop automatic ways to generate narratives by combining short video segments and tailoring this to the learner’s needs. As a first step, this paper is proposing an original framework to characterise video segments for learning by combining video content and audience attention. The input for the framework includes the video transcripts, past user interactions with the videos, and an ontology defining the core domain concepts. The output is a set of patterns that are associated with the video segments, describing the focus topic and concepts of the segment. We have applied the framework on a dataset from user studies with the AVW space for presentation skills learning, including 49 video segments that are high attention intervals from past user interactions. The video segment characterisation provides useful insights to inform recommendations and segment combinations to support informal learning.