Research

Paper

AI LLM March 25, 2026

ConceptKT: A Benchmark for Concept-Level Deficiency Prediction in Knowledge Tracing

Authors

Yu-Chen Kang, Yu-Chien Tang, An-Zi Yen

Abstract

Knowledge Tracing (KT) is a critical technique for modeling student knowledge to support personalized learning. However, most KT systems focus on binary correctness prediction and cannot diagnose the underlying conceptual misunderstandings that lead to errors. Such fine-grained diagnostic feedback is essential for designing targeted instruction and effective remediation. In this work, we introduce the task of concept-level deficiency prediction, which extends traditional KT by identifying the specific concepts a student is likely to struggle with on future problems. We present ConceptKT, a dataset annotated with labels that capture both the concepts required to solve each question and the missing concepts underlying incorrect responses. We investigate in-context learning approaches to KT and evaluate the diagnostic capabilities of various Large Language Models (LLMs) and Large Reasoning Models (LRMs). Different strategies for selecting informative historical records are explored. Experimental results demonstrate that selecting response histories based on conceptual alignment and semantic similarity leads to improved performance on both correctness prediction and concept-level deficiency identification.

Metadata

arXiv ID: 2603.24073
Provider: ARXIV
Primary Category: cs.CL
Published: 2026-03-25
Fetched: 2026-03-26 06:02

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.24073v1</id>\n    <title>ConceptKT: A Benchmark for Concept-Level Deficiency Prediction in Knowledge Tracing</title>\n    <updated>2026-03-25T08:27:01Z</updated>\n    <link href='https://arxiv.org/abs/2603.24073v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.24073v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Knowledge Tracing (KT) is a critical technique for modeling student knowledge to support personalized learning. However, most KT systems focus on binary correctness prediction and cannot diagnose the underlying conceptual misunderstandings that lead to errors. Such fine-grained diagnostic feedback is essential for designing targeted instruction and effective remediation. In this work, we introduce the task of concept-level deficiency prediction, which extends traditional KT by identifying the specific concepts a student is likely to struggle with on future problems. We present ConceptKT, a dataset annotated with labels that capture both the concepts required to solve each question and the missing concepts underlying incorrect responses. We investigate in-context learning approaches to KT and evaluate the diagnostic capabilities of various Large Language Models (LLMs) and Large Reasoning Models (LRMs). Different strategies for selecting informative historical records are explored. Experimental results demonstrate that selecting response histories based on conceptual alignment and semantic similarity leads to improved performance on both correctness prediction and concept-level deficiency identification.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CL'/>\n    <published>2026-03-25T08:27:01Z</published>\n    <arxiv:comment>Accepted by LREC 2026</arxiv:comment>\n    <arxiv:primary_category term='cs.CL'/>\n    <author>\n      <name>Yu-Chen Kang</name>\n    </author>\n    <author>\n      <name>Yu-Chien Tang</name>\n    </author>\n    <author>\n      <name>An-Zi Yen</name>\n    </author>\n  </entry>"
}