MAML-KT: Addressing Cold Start Problem in Knowledge Tracing for New Students via Few-Shot Model-Agnostic Meta Learning

Researchers introduce MAML-KT, a meta-learning approach that addresses the cold start problem in knowledge tracing systems when predicting performance of new students with limited interaction data. The model uses few-shot learning to rapidly adapt to unseen students, achieving higher early accuracy than existing knowledge tracing models across multiple datasets.

• →MAML-KT addresses the cold start problem in educational AI by using meta-learning to predict new student performance with minimal data.
• →The approach outperforms existing knowledge tracing models like DKT, DKVMN, and SAKT in early prediction accuracy for new students.
• →Testing across ASSIST datasets shows consistent improvements in cold start scenarios with cohorts of 10-50 students.
• →Performance drops correlate with skill novelty rather than model instability, providing clearer interpretation of learning dynamics.
• →The research highlights limitations of standard evaluation methods that obscure real-world deployment challenges in educational technology.