Synthesis
Novel generate-then-validate approach for educational question generation. Reduces hallucination by 62% compared to direct generation. Validated on STEM datasets with 89% accuracy. Outperforms baseline LLMs by 23% on relevance metrics.
Connections
- automated-question-generation — Core methodology contribution
- nlp-education — LLM application for assessment
- stem-education — Primary validation domain
- llm-application-education — Generative AI for assessment
- automated-assessment — Reduces manual question authoring burden
- educational-nlp — NLP techniques for learning materials
Related Pages
- slidesqaqa-pedagogical-question-generation — contrasts with front-loaded pedagogical reasoning approach
- short-answer-scoring-quality-degradation — Quality assurance methodology for automated assessment generation
- nsmq-riddles-science-math-benchmark — Educational question benchmark from the Global South
References
Wei, Y., Stamper, J., et al. (2025). Generate-Then-Validate: A Novel Question Generation Approach. arXiv preprint arXiv:2512.10110.