![[2602.20135] KNIGHT: Knowledge Graph-Driven Multiple-Choice Question Generation with Adaptive Hardness Calibration](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.20135] KNIGHT: Knowledge Graph-Driven Multiple-Choice Question Generation with Adaptive Hardness Calibration
Summary
The paper introduces KNIGHT, a framework for generating multiple-choice questions using knowledge graphs, enhancing efficiency and adaptability in educational assessments.
Why It Matters
As large language models become prevalent in educational technology, KNIGHT addresses the challenge of creating assessment datasets efficiently. By leveraging knowledge graphs, it offers a scalable solution for generating high-quality, difficulty-adjusted questions, which can significantly aid educators and researchers in evaluating learning outcomes.
Key Takeaways
- KNIGHT utilizes knowledge graphs to streamline the generation of multiple-choice questions.
- The framework supports adjustable difficulty levels for tailored assessments.
- Quality evaluation metrics include fluency, topic relevance, and answerability.
- KNIGHT demonstrates cost-efficiency in generating educational content.
- The approach is domain-agnostic, applicable across various subjects.
Computer Science > Computation and Language arXiv:2602.20135 (cs) [Submitted on 23 Feb 2026] Title:KNIGHT: Knowledge Graph-Driven Multiple-Choice Question Generation with Adaptive Hardness Calibration Authors:Mohammad Amanlou, Erfan Shafiee Moghaddam, Yasaman Amou Jafari, Mahdi Noori, Farhan Farsi, Behnam Bahrak View a PDF of the paper titled KNIGHT: Knowledge Graph-Driven Multiple-Choice Question Generation with Adaptive Hardness Calibration, by Mohammad Amanlou and 5 other authors View PDF HTML (experimental) Abstract:With the rise of large language models (LLMs), they have become instrumental in applications such as Retrieval-Augmented Generation (RAG). Yet evaluating these systems remains bottlenecked by the time and cost of building specialized assessment datasets. We introduce KNIGHT, an LLM-based, knowledge-graph-driven framework for generating multiple-choice question (MCQ) datasets from external sources. KNIGHT constructs a topic-specific knowledge graph, a structured and parsimonious summary of entities and relations, that can be reused to generate instructor-controlled difficulty levels, including multi-hop questions, without repeatedly re-feeding the full source text. This knowledge graph acts as a compressed, reusable state, making question generation a cheap read over the graph. We instantiate KNIGHT on Wikipedia/Wikidata while keeping the framework domain- and ontology-agnostic. As a case study, KNIGHT produces six MCQ datasets in History, Biology, and Mathe...
