Uncovering patterns in student work : machine learning to understand human learning
Abstract/Contents
- Abstract
- When millions of students learn by working through the same curriculum what patterns unfold? Being able to autonomously understand students learning, both in terms of assessing knowledge and being able to provide feedback, is a grand challenge in education and one that has been studied for many years. Massive online classes, which were recently popularized, have provided a serendipitous opportunity to break ground on this important problem. In my dissertation I will explore data driven approaches in the domain of students learning to program -- a discipline with rich, structured assignments. I will discuss emergent patterns in; the space of student partial solutions, in how students navigate open-ended assignments and in how students work through a series of problems. Highlights include (1) the first application of deep learning to student trajectories which yields a noteworthy improvement in the state of the art for the task of Knowledge Tracing (2) discovery of a pattern in how students navigate solution spaces that both predicts how teachers would suggest a learner make forward progress and has an almost perfect logarithmic relationship with the probability of a student succeeding in the future and (3) A method to apply deep neural networks to autonomously embed student programs into Euclidian space. This work has been in collaboration with some of the largest online platforms to date: Khan Academy, Coursera and Code.org. My thesis is meant to depict a nascent subspace of education research, and where it could go. The results I present more than anything show the extent to which many open, fascinating problems remain.
Description
| Type of resource | text |
|---|---|
| Form | electronic; electronic resource; remote |
| Extent | 1 online resource. |
| Publication date | 2016 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Associated with | Piech, Christopher James |
|---|---|
| Associated with | Stanford University, Department of Computer Science. |
| Primary advisor | Guibas, Leonidas J |
| Thesis advisor | Guibas, Leonidas J |
| Thesis advisor | Mitchell, John |
| Thesis advisor | Sahami, Mehran |
| Advisor | Mitchell, John |
| Advisor | Sahami, Mehran |
Subjects
| Genre | Theses |
|---|
Bibliographic information
| Statement of responsibility | Christopher James Piech. |
|---|---|
| Note | Submitted to the Department of Computer Science. |
| Thesis | Thesis (Ph.D.)--Stanford University, 2016. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2016 by Christopher James Piech
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...