KASER : Knowledge-Aligned Student Error Simulator for Open-Ended Coding Tasks
Zhangqi Duan, Nigel Fernandez, Andrew Lan · 2026
Read KASER : Knowledge-Aligned Student Error Simulator for Open-Ended Coding Tasks by Zhangqi Duan, Nigel Fernandez, Andrew Lan (2026) 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-2601-06633.txt.
Description
Zhangqi Duan, Nigel Fernandez, and Andrew Lan present KASER in 2026. KASER simulates student errors on open-ended coding tasks for education. Such tasks give detailed insights into student knowledge, they note.
Training language models to simulate student errors is challenging, though. Models often suffer mode collapse when generating plausible wrong answers. They also miss diversity in syntax, style, and solution approach. Student responses vary widely on all three of those axes. KASER is knowledge-aligned, aiming to match errors to student understanding. Better simulators could help teachers anticipate mistakes before grading piles arrive. Computing education readers get a tool for generating realistic wrong answers. That supports practice, feedback, and test design at scale.
Catalog record
| CREATOR | Zhangqi Duan, Nigel Fernandez, Andrew Lan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 59 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-06633