Research map / Vision-language and multimodal / Visual question answering
Medical Exam Question Answering with Large-scale Reading Comprehension
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use SeaReader instead of R-NET | MedQA · MedQA test set | accuracy: 64.5 -> 73.6 | Helped |
| Ensemble SeaReader models | SeaReader · MedQA test set | accuracy: 73.6 -> 75.3 | Helped |
| Increase retrieved documents per candidate answer from top-1 to top-5 | SeaReader · MedQA test set | accuracy: 57.8 -> 71.7 | Helped |
| Increase retrieved documents per candidate answer from top-5 to top-10 | SeaReader · MedQA test set | accuracy: 71.7 -> 73.6 | Helped |
| Increase retrieved documents per candidate answer from top-10 to top-20 | SeaReader · MedQA test set | accuracy: 73.6 -> 74.4 | Helped |
| Add a context BiLSTM instead of removing the context layer (ablation) | SeaReader · MedQA | accuracy | Helped |
| Use a context BiLSTM instead of a convolutional context layer (ablation) | SeaReader · MedQA | accuracy | Helped |
| Use a context BiLSTM instead of a GRU context layer (ablation) | SeaReader · MedQA | accuracy | Helped |
| Use delta embeddings with L21 regularization instead of directly learning embeddings (ablation) | SeaReader · MedQA | accuracy | Helped |
| Adapt pretrained embeddings rather than keeping them fixed (ablation) | SeaReader · MedQA | accuracy | Helped |
| Use max- and mean-pooling matching features instead of CNN matching features (ablation) | SeaReader · MedQA | accuracy | No clear effect |
| Add noisy gating (ablation) | SeaReader · MedQA | generalization performance | Helped |
Closest papers by meaning
- NovelQA: Benchmarking Question Answering on Documents Exceeding 200K Tokens (ICLR 2025 · similarity 0.87)
- Detecting Any instruction-to-answer interaction relationship:Universal Instruction-to-Answer Navigator for Med-VQA (ICML 2024 · similarity 0.85)
- DiN: Diffusion Model for Robust Medical VQA with Semantic Noisy Labels (CVPR 2025 · similarity 0.85)
- Visual and Domain Knowledge for Professional-level Graph-of-Thought Medical Reasoning (ICML 2025 · similarity 0.85)
- Falcon Medical Visual Question Answering (AAAI 2025 · similarity 0.85)
- DSAS: A Universal Plug-and-Play Framework for Attention Optimization in Multi-Document Question Answering (NeurIPS 2025 · similarity 0.85)
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.