mlsanityData to 14 Oct 2026

Candidate-Expanding Routing with Permutation-Stabilized Experts for Mixed-Format Medical VQA

arXiv 2609.00959 ↗First posted 1 Sep 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Mixed-format medical visual question answering (VQA) requires stable option selection and machine-readable free-text output. The two formats fail differently: multiple-choice predictions can change with option symbols or positions, while clinically plausible open answers can fail automated evaluation when serialization is malformed. We address both challenges with an answer-text memory, a permutation-stabilized vision--language expert, and a sparse candidate- expanding router. The cyclic schedule follows prior work; our contribution is to make expert top-2 a routable candidate alongside memory and expert top-1. On a 1,403-case retrospective internal analysis, this expansion improves a matched binary router from 88.95% to 91.73% (+2.78 percentage points; 95% CI 1.57--3.99), with 56 rescued errors and 17 regressions. Oracle coverage rises from 90.31% to 96.15%, and the final submitted configuration reaches 92.23% on the same retrospective split. For open questions, strict generation and deterministic guards produce 475/475 schema- valid participant-facing outputs without repair, retry, or hard-gate failure. Visual ablations reveal substantial textual dependence. Candidate expansion supplies the principal controlled routing gain; open-path evidence establishes output-contract validity rather than clinical correctness in medical use or deployment.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 21 of its 26 references that are in the map. First posted on arXiv 1 Sep 2026.

Builds on

21 of 26 references.

References · 21

  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. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  4. 2025Qwen2.5-VL Technical Reportcited by 5,587
  5. 2021Calibrate Before Use: Improving Few-Shot Performance of Language Modelscited by 1,392
  6. 2023LLaVA-Med: Training a Large Language-and-Vision Assistant for Biomedicine in One Daycited by 1,305
  7. 2017Selective Classification for Deep Neural Networkscited by 852
  8. 2020Towards Faithfully Interpretable NLP Systems: How should we define and evaluate faithfulness?cited by 599
  9. 2023Large Language Models Are Not Robust Multiple Choice Selectorscited by 486
  10. 2025MedGemma Technical Reportcited by 436
  11. 2023PMC-VQA: Visual Instruction Tuning for Medical Visual Question Answeringcited by 338
  12. 2020Consistent Estimators for Learning to Defer to an Expertcited by 271
9 more
  1. 2023Large Language Models Sensitivity to The Order of Options in Multiple-Choice Questionscited by 263
  2. 2025Lingshu: A Generalist Foundation Model for Unified Multimodal Medical Understanding and Reasoningcited by 230
  3. 2021Medical Visual Question Answering: A Surveycited by 93
  4. 2023RAMM: Retrieval-augmented Biomedical Visual Question Answering with Multi-modal Pre-trainingcited by 31
  5. 2023Fool Your (Vision and) Language Model With Embarrassingly Simple Permutationscited by 25
  6. 2024MedVH: Towards Systematic Evaluation of Hallucination for Large Vision Language Models in the Medical Contextcited by 20
  7. 2025Mastering Multiple-Expert Routing: Realizable H-Consistency and Strong Guarantees for Learning to Defercited by 16
  8. 2025Med-CMR: A Fine-Grained Benchmark Integrating Visual Evidence and Clinical Logic for Medical Complex Multimodal Reasoningcited by 14
  9. 2025Benchmarking and Mitigating MCQA Selection Bias of Large Vision-Language Modelscited by 5

Built on by

Not cited yet.