Knowledge Tracing Models: Predictive Accuracy vs. Parameter Recovery Explored

A new research paper published on arXiv explores the nuances of knowledge tracing models, specifically focusing on whether predictive accuracy equates to accurate recovery of item parameters. The study investigates various decoder and encoder architectures, including LSTM, Transformer, and DKVMN, and finds that while difficulty and intercept parameters are recovered well, discrimination and option slope recovery are more sensitive to representation width. The research suggests that separating item parameter keys from sequence representations can improve recovery and consistency without sacrificing predictive accuracy. AI

IMPACT This research clarifies how to better evaluate the internal parameters of knowledge tracing models, potentially leading to more reliable educational AI tools.

RANK_REASON The cluster contains a research paper detailing a study on knowledge tracing models. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• arXiv
• DKVMN
• Irt
• LSTM
• Transformer

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →