Research map / Large language models / Retrieval and knowledge
On the Generation of Medical Question-Answer Pairs
Sheng Shen, Yaliang Li, Nan Du, Xian Wu, Yusheng Xie, Shen Ge, Tao Yang, Kai Wang, Xingzheng Liang, Wei Fan
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add generated QA pairs from eg-CVAE to the training set | NMLEC examination QA system · NMLEC_2017 | Accuracy: 61.97 -> 67.96 | Improved |
| add generated QA pairs from HRED to the training set | NMLEC examination QA system · NMLEC_2017 | Accuracy: 61.97 -> 58.78 | Hurt |
| add generated QA pairs from VHRED to the training set | NMLEC examination QA system · NMLEC_2017 | Accuracy: 61.97 -> 62.28 | Improved |
| add type-level modeling and retain type-only decoding (ablation) | eg-CVAE · NMLEC_2017 | Accuracy: 61.97 -> 65.27 | Improved |
| add entity-level modeling and retain entity-only decoding (ablation) | eg-CVAE · NMLEC_2017 | Accuracy: 61.97 -> 64.67 | Improved |
| combine type and entity decoding in the multi-pass eg-CVAE | NMLEC medical question generation · NMLEC_QA generated questions | inter-dist-1: 0.533 -> 0.637 | Improved |
| replace combined type-and-entity passes with type-only decoding (ablation) | eg-CVAE · NMLEC_QA generated questions | BLEU F1: 0.494 -> 0.572 | Improved |
| replace combined type-and-entity passes with entity-only decoding (ablation) | eg-CVAE · NMLEC_QA generated questions | BLEU F1: 0.494 -> 0.613 | Improved |
| use type-only rather than combined decoding (ablation) | eg-CVAE · NMLEC_QA generated questions | intra-dist-2: 0.994 -> 0.997 | Improved |
| use entity-only rather than combined decoding (ablation) | eg-CVAE · NMLEC_QA generated questions | BLEU Precision: 0.450 -> 0.541 | Improved |
| combine implicit phrase-type and explicit entity constraints | medical QA generation · NMLEC_QA human evaluation | Consistency: 3.68 -> 4.09 | Improved |
| combine implicit phrase-type and explicit entity constraints | medical QA generation · NMLEC_QA human evaluation | Fluency: 4.08 -> 4.43 | Improved |
Closest papers by meaning
- MedAlign: A Clinician-Generated Dataset for Instruction Following with Electronic Medical Records (AAAI 2024 · similarity 0.85)
- QGen Studio: An Adaptive Question-Answer Generation, Training and Evaluation Platform (AAAI 2025 · similarity 0.85)
- EHRNote-ChatQA: A Benchmark for Evidence-Grounded Multi-Turn Clinical Question Answering over Longitudinal Discharge Summaries (NeurIPS 2026 · similarity 0.84)
- Falcon Medical Visual Question Answering (AAAI 2025 · similarity 0.84)
- Achieving Human Parity in Content-Grounded Datasets Generation (ICLR 2024 · similarity 0.83)
- Alignment, Mining and Fusion: Representation Alignment with Hard Negative Mining and Selective Knowledge Fusion for Medical Visual Question Answering (CVPR 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
