Technology Category
Knowledge Representation and Reasoning: Knowledge AcquisitionData Mining & Knowledge Management: Knowledge Acquisition from the WebCognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
PICKT: Practical Interlinked Concept Knowledge Tracing for Personalized Learning using Knowledge Map Concept Relations
🤖 AI Summary
Existing knowledge tracing (KT) models face critical challenges including rigid input formats, cold-start issues for new students and new items, and insufficient online service stability. To address these, we propose a dynamic KT framework that jointly leverages knowledge graphs and textual semantics. First, we construct a concept-aware knowledge graph encoding semantic relationships among domain concepts, enabling flexible integration of heterogeneous inputs. Second, we design an attention-based concept relation aggregation network that jointly models sequential learning behaviors and the underlying conceptual structure. Notably, our approach is the first to systematically tackle the dual cold-start problem in KT—simultaneously for unseen students and unseen questions. Extensive experiments on real-world educational datasets demonstrate significant improvements in prediction accuracy for both new students and new items. Moreover, our method outperforms state-of-the-art baselines in stability, generalizability, and practical deployability, validating its robustness and operational value in complex, production-grade educational platforms.
Technology Category
Application Category
📝 Abstract
With the recent surge in personalized learning, Intelligent Tutoring Systems (ITS) that can accurately track students' individual knowledge states and provide tailored learning paths based on this information are in demand as an essential task. This paper focuses on the core technology of Knowledge Tracing (KT) models that analyze students' sequences of interactions to predict their knowledge acquisition levels. However, existing KT models suffer from limitations such as restricted input data formats, cold start problems arising with new student enrollment or new question addition, and insufficient stability in real-world service environments. To overcome these limitations, a Practical Interlinked Concept Knowledge Tracing (PICKT) model that can effectively process multiple types of input data is proposed. Specifically, a knowledge map structures the relationships among concepts considering the question and concept text information, thereby enabling effective knowledge tracing even in cold start situations. Experiments reflecting real operational environments demonstrated the model's excellent performance and practicality. The main contributions of this research are as follows. First, a model architecture that effectively utilizes diverse data formats is presented. Second, significant performance improvements are achieved over existing models for two core cold start challenges: new student enrollment and new question addition. Third, the model's stability and practicality are validated through delicate experimental design, enhancing its applicability in real-world product environments. This provides a crucial theoretical and technical foundation for the practical implementation of next-generation ITS.
Problem
Research questions and friction points this paper is trying to address.
Overcomes cold start issues in knowledge tracing models
Enhances stability and practicality for real-world tutoring systems
Utilizes knowledge maps to process diverse input data effectively
Innovation
Methods, ideas, or system contributions that make the work stand out.
Knowledge map structures concept relationships for tracing
Model processes multiple data types to overcome limitations
Enhances stability and performance in cold start scenarios
🔎 Similar Papers
No similar papers found.
