Date
AI Overview
Why it matters
Real-world K-12 learning often uses multimodal materials with long student histories. Existing AI tools typically focus on plain text and single tasks, which limits generalization and incurs high computing costs.
Study at a glance
- Question: Can a unified multimodal model handle multiple educational tasks efficiently while compressing long student interaction histories?
- Setting and data: Exercise platform records from 13,239 students, covering 8,247 math concepts, 235,687 questions, and 3,892,084 interactions.
- Method: Machine learning model development and empirical benchmark evaluation across four educational tasks.
- AI tools: UniEDU, Qwen2-VL-2B, Qwen2-VL-7B, Qwen2.5-0.5B, Qwen2.5-1.5B, Qwen2.5-3B
Key findings
- UniEDU surpassed task-specific baselines in knowledge recommendation, time cost prediction, and answer prediction, gaining roughly 30% and 20% over top baselines on the first two.
- Compressing student interaction histories reduced memory overhead by over three times compared to a smaller baseline model, while preserving competitive task accuracy.
- Larger language model sizes improved performance on complex generative tasks such as user answer prediction, while simpler discriminative tasks showed minimal gains.
Summary drafted with AI from the full paper and may contain errors. See the full paper for details.
Education materials for K-12 students often consist of multiple modalities,
such as text and images, posing challenges for models to fully understand
nuanced information in these materials. In this paper, we propose a unified
language and vision assistant UniEDU designed for various educational
applications, including knowledge recommendation, knowledge tracing, time cost
prediction, and user answer prediction, all within a single model. Unlike
conventional task-specific models, UniEDU offers a unified solution that excels
across multiple educational tasks while maintaining strong generalization
capabilities. Its adaptability makes it well-suited for real-world deployment
in diverse learning environments. Furthermore, UniEDU is optimized for
industry-scale deployment by significantly reducing computational
overhead-achieving approximately a 300\% increase in efficiency-while
maintaining competitive performance with minimal degradation compared to fully
fine-tuned models. This work represents a significant step toward creating
versatile AI systems tailored to the evolving demands of education.
Research Categories
What is the application?
Who is the user?
Why use AI?
Study design
AI tools
Publication Source
Publisher
arXiv
