Disentangled Knowledge Tracing for Alleviating Cognitive Bias
Yiyun Zhou, Zheqi Lv, Shengyu Zhang, Jingyuan Chen · 2025
Read Disentangled Knowledge Tracing for Alleviating Cognitive Bias by Yiyun Zhou, Zheqi Lv, Shengyu Zhang, Jingyuan Chen (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-02539.txt.
Description
Yiyun Zhou, Zheqi Lv, Shengyu Zhang, and Jingyuan Chen aim to free tutoring software from a bias that skews how it judges students. Their paper appeared in two thousand twenty-five. Such software tracks what each student knows to pick the next task. When some topics appear far more than others, strong students get too little work and weak ones get too much.
The team traced the skew to a hidden factor: each student's past success rate across topic groups. Their model, DisKT, splits what a student knows into familiar and new ground. It then strips the hidden factor out of the student profile. A further part shields the model from guesses and slips in the data. Tests on many data sets show less bias and sharper calls than past tools. The model's picks can also be read in plain terms.
Catalog record
| CREATOR | Yiyun Zhou, Zheqi Lv, Shengyu Zhang, Jingyuan Chen |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 66 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-02539