mlsanityData to 14 Oct 2026

Option-Aware Retrieval and Task-Specific VLM Adaptation for Medical VQA

arXiv 2609.15530 ↗First posted 14 Sep 2026cs.AI cs.CL cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

We describe our submission to the MedReason 2026 challenge, covering multiple-choice (MCQ) and open-ended (OE) medical visual question answering (VQA) under fully offline, containerized inference. Our first finding is that MCQ retrieval must compare answer semantics rather than answer labels: labels are independently assigned per question, so copying a retrieved neighbor's label transfers no useful information, whereas scoring each current option's text against correct-answer text from similar training cases raises retrieval-only accuracy from 20.0% to 57.5% on a 200-case retrieval-excluded development holdout. Our second finding attributes the submitted system's accuracy: holding the task-specific MCQ Low-Rank Adaptation (LoRA) adapter fixed and varying the number (k) of in-prompt retrieved examples changes accuracy by at most one case --- 187/200 (93.5%) at both (k=0) and the adapter's training-time (k=1), 188/200 (94.0%) at the packaged runtime's default (k=3) --- and the submitted confidence-gated override adds no net accuracy on top of (k=3), selecting the VLM in 198/200 cases. With the final MCQ adapter fixed, retrieval changes accuracy by at most one case, and gating provides no net gain. On 20 OE cases, token-F1 and RaTEScore~citezhao2024ratescore decrease as (k) grows, but paired sign tests on token-F1 differences are nonsignificant ((p ≥ 0.29)); a single-annotator comparison found 6/20 wrong-anchor errors for the final configuration and 14/20 for an earlier configuration that jointly differed in routing, adapter, and prompting. The system reaches 94.0% MCQ accuracy on the development holdout and 93.20% on the organizer's official pre-evaluation, versus 29.43% for the off-the-shelf reference baseline, while both of the organizer's open-ended scores are lower than that baseline's (ground-truth agreement 1.245 versus 1.588, visual accuracy 1.995 versus 2.696, each out of 4).

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 10 of its 14 references that are in the map. First posted on arXiv 14 Sep 2026.

Builds on

10 of 14 references.

References · 10

  1. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  2. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  3. 2025Qwen2.5-VL Technical Reportcited by 5,587
  4. 2017On Calibration of Modern Neural Networkscited by 4,960
  5. 2023LLaVA-Med: Training a Large Language-and-Vision Assistant for Biomedicine in One Daycited by 1,305
  6. 2021SLAKE: A Semantically-Labeled Knowledge-Enhanced Dataset for Medical Visual Question Answeringcited by 449
  7. 2023PMC-VQA: Visual Instruction Tuning for Medical Visual Question Answeringcited by 338
  8. 2024Benchmarking Retrieval-Augmented Generation for Medicinecited by 331
  9. 2024RaTEScore: A Metric for Radiology Report Generationcited by 65
  10. 2025Med-CMR: A Fine-Grained Benchmark Integrating Visual Evidence and Clinical Logic for Medical Complex Multimodal Reasoningcited by 14

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026MedProb: Probing Internal Representations of Vision-Language Models for Medical Question Answering
  2. 2025DiN: Diffusion Model for Robust Medical VQA with Semantic Noisy Labels
  3. 2026LiteMedCoT-VL: Parameter-Efficient Adaptation for Medical Visual Question Answering
  4. 2025CoTBox-TTT: Grounding Medical VQA with Visual Chain-of-Thought Boxes During Test-time Training
  5. 2025Synthesizing High-Quality Visual Question Answering from Medical Documents with Generator-Verifier LMMs
  6. 2023Visual Question Answering in the Medical Domain
  7. 2024MISS: A Generative Pretraining and Finetuning Approach for Med-VQA
  8. 2025Multimodal Retrieval-Augmented Generation with Large Language Models for Medical VQA