2020/07/13 by Anirudh Som, Som, Anirudh, Sujeong Kim +9
Computer Science · #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Online Learning and Analytics #Text and Document Classification Technologies
paper · pdf · doi:10.48550/arxiv.2007.06667
openalex publication_date 2020/07/13 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
K-12 classrooms consistently integrate collaboration as part of their\nlearning experiences. However, owing to large classroom sizes, teachers do not\nhave the time to properly assess each student and give them feedback. In this\npaper we propose using simple deep-learning-based machine learning models to\nautomatically determine the overall collaboration quality of a group based on\nannotations of individual roles and individual level behavior of all the\nstudents in the group. We come across the following challenges when building\nthese models: 1) Limited training data, 2) Severe class label imbalance. We\naddress these challenges by using a controlled variant of Mixup data\naugmentation, a method for generating additional data samples by linearly\ncombining different pairs of data samples and their corresponding class labels.\nAdditionally, the label space for our problem exhibits an ordered structure. We\ntake advantage of this fact and also explore using an ordinal-cross-entropy\nloss function and study its effects with and without Mixup.\n