Agentic data science with the Data Agent Kit
cloud.google.com·6h ago
AI Research
arXiv CS.AI·Aug 26·5 min read
Incorporating Cognitive Load and Knowledge Transfer for Multi-Domain Knowledge Tracing
TL;DR
A novel method called LT-MKT enhances multi-domain knowledge tracing by integrating cognitive load and knowledge transfer, leading to improved predictions of student performance. Extensive experiments show it achieves state-of-the-art results on real-world datasets.
✦ Why It Matters
Implement LT-MKT in your educational tools to enhance student performance predictions across multiple subjects.
Key Takeaways
- Integrate LT-MKT for better multi-domain learning assessments.
- Utilize large language models for knowledge representation.
- Model cognitive load to improve learning predictions.
- Track knowledge transfer between domains for enhanced insights.
How It Works
LT-MKT constructs a Multi-domain Hierarchical Graph by integrating textual information from questions and concepts, leveraging large language models for advanced representation. It models cognitive load by capturing cross-domain features in both temporal and knowledge dimensions, while a dedicated knowledge transfer module tracks how knowledge states influence each other across different domains.