Research map / Vision-language and multimodal / Visual question answering
EHRXQA: A Multi-Modal Question Answering Dataset for Electronic Health Records with Chest X-ray Images
Seongsu Bae, Daeun Kyung, Jaehee Ryu, Eunbyeol Cho, Gyubok Lee, Sunjun Kweon, Jungwoo Oh, Lei Ji, Eric I-Chao Chang, Tackeun Kim, Edward Choi
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 |
|---|---|---|---|
| Pre-train the PubMedCLIP model on the shared MIMIC-CXR image-report corpus | PubMedCLIP on MIMIC-CXR-VQA · MIMIC-CXR-VQA test | accuracy: $54.9 \pm 1.3$ -> $56.5 \pm 2.1$ | Improved |
| Pre-train MedViLL on the shared MIMIC-CXR image-report corpus | MedViLL on MIMIC-CXR-VQA · MIMIC-CXR-VQA test | accuracy: 63.6 -> 63.6 | Improved |
| Pre-train M³AE on the shared MIMIC-CXR image-report corpus | M³AE on MIMIC-CXR-VQA · MIMIC-CXR-VQA test | accuracy: $68.9 \pm 0.3$ -> $69.2 \pm 0.4$ | Improved |
| Retrieve question-specific few-shot examples with BM25 instead of using a fixed prompt | ChatGPT + M³AE on EHRXQA · Image-related QA | Acc_EX|pred: 17.4 -> 48.2 | Improved |
| Retrieve question-specific few-shot examples with BM25 instead of using a fixed prompt | ChatGPT + M³AE on EHRXQA · Table-related QA | Acc_EX|pred: 30.0 -> 92.9 | Improved |
| Retrieve question-specific few-shot examples with BM25 instead of using a fixed prompt | ChatGPT + M³AE on EHRXQA · Image+Table-related QA | Acc_EX|pred: 35.7 -> 65.9 | Improved |
| Require the VQA model to process multiple images for a single-patient question | ChatGPT + M³AE on EHRXQA · Image-related, single-patient N-image scope | Acc_EX|pred: 48.2 -> 39.6 | Hurt |
| Expand image-based QA from a patient scope to a group scope | ChatGPT + M³AE on EHRXQA · Image-related group scope | Acc_EX|pred: 48.2 -> 1.7 | Hurt |
| Constrain the image scope with table conditions in combined-modality questions | ChatGPT + M³AE on EHRXQA · Image+Table-related versus Image-related QA | Acc_EX|pred: 48.2 -> 65.9 | Improved |
Closest papers by meaning
- Alignment, Mining and Fusion: Representation Alignment with Hard Negative Mining and Selective Knowledge Fusion for Medical Visual Question Answering (CVPR 2025 · similarity 0.89)
- MedAtlas: Evaluating LLMs for Multi-Round, Multi-Task Medical Reasoning Across Diverse Imaging Modalities and Clinical Text (AAAI 2026 · similarity 0.89)
- Falcon Medical Visual Question Answering (AAAI 2025 · similarity 0.88)
- OmniMedVQA: A New Large-Scale Comprehensive Evaluation Benchmark for Medical LVLM (CVPR 2024 · similarity 0.88)
- DiN: Diffusion Model for Robust Medical VQA with Semantic Noisy Labels (CVPR 2025 · similarity 0.88)
- 3D-RAD: A Comprehensive 3D Radiology Med-VQA Dataset with Multi-Temporal Analysis and Diverse Diagnostic Tasks (NeurIPS 2025 · similarity 0.87)
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.
