Difficulty-Controllable Multiple-Choice Question Generation Using Large Language Models and Direct Preference Optimization

2025-10-22

AI Summary

Existing controllable difficulty generation methods for reading comprehension questions struggle to directly produce mainstream educational multiple-choice questions (MCQs) and lack fine-grained difficulty control. Method: This paper proposes an end-to-end, difficulty-controllable MCQ generation framework leveraging large language models (LLMs) and Direct Preference Optimization (DPO). It introduces differentiable difficulty prompts and jointly models explicit difficulty levels with structured distractor generation, enabling preference-aligned optimization over human-annotated difficulty preferences. Contribution/Results: Our approach achieves the first explicit difficulty-level modeling and simultaneous option structure generation for MCQs. Experiments across multiple educational benchmarks demonstrate significant improvements over baselines in difficulty adherence rate, question quality, and answer-option diversity. The method supports personalized, adaptive assessment and exhibits strong practical viability for real-world educational deployment.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Natural Language Processing: Generation
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Difficulty-controllable question generation for reading comprehension has gained significant attention in the field of education as a fundamental tool for adaptive learning support. Although several neural question generation methods have recently succeeded in controlling difficulty, conventional approaches still face two major limitations. First, they cannot directly generate multiple-choice questions, which are the most widely used question type in educational contexts. Second, they are not explicitly trained to optimize the accuracy of difficulty control, leaving room for further improvement in difficulty controllability. To address these limitations, this study proposes a novel difficulty-controllable multiple-choice question generation method for reading comprehension which leverages a large language model trained using a direct preference optimization technique to improve the accuracy of difficulty control.

Problem

• Generating difficulty-controllable multiple-choice reading comprehension questions
• Overcoming limitations of conventional neural question generation methods
• Improving difficulty control accuracy via direct preference optimization

Innovation

• Uses large language models for question generation
• Applies direct preference optimization for difficulty control
• Generates multiple-choice questions with controllable difficulty