We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CY

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computers and Society

Title: Meta Transfer Learning for Early Success Prediction in MOOCs

Abstract: Despite the increasing popularity of massive open online courses (MOOCs), many suffer from high dropout and low success rates. Early prediction of student success for targeted intervention is therefore essential to ensure no student is left behind in a course. There exists a large body of research in success prediction for MOOCs, focusing mainly on training models from scratch for individual courses. This setting is impractical in early success prediction as the performance of a student is only known at the end of the course. In this paper, we aim to create early success prediction models that can be transferred between MOOCs from different domains and topics. To do so, we present three novel strategies for transfer: 1) pre-training a model on a large set of diverse courses, 2) leveraging the pre-trained model by including meta information about courses, and 3) fine-tuning the model on previous course iterations. Our experiments on 26 MOOCs with over 145,000 combined enrollments and millions of interactions show that models combining interaction data and course information have comparable or better performance than models which have access to previous iterations of the course. With these models, we aim to effectively enable educators to warm-start their predictions for new and ongoing courses.
Comments: Accepted at the 2022 ACM Conference on Learning at Scale (L@S 2022)
Subjects: Computers and Society (cs.CY); Machine Learning (cs.LG)
Cite as: arXiv:2205.01064 [cs.CY]
  (or arXiv:2205.01064v1 [cs.CY] for this version)

Submission history

From: Vinitra Swamy [view email]
[v1] Mon, 25 Apr 2022 09:19:57 GMT (26957kb,D)

Link back to: arXiv, form interface, contact.