Skip to content

"Worse than Random? An Embarrassingly Simple Probing Evaluation of Large Multimodal Models in Medical VQA"

Notifications You must be signed in to change notification settings

eric-ai-lab/ProbMed

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

38 Commits
 
 
 
 
 
 

Repository files navigation

ProbMed

🌐 Homepage | 🤗 Dataset | 🤗 Paper | 📖 arXiv | GitHub

This repo contains the evaluation code for the paper "Worse than Random? An Embarrassingly Simple Probing Evaluation of Large Multimodal Models in Medical VQA"

Introduction

We introduce the Probing Evaluation for Medical Diagnosis (ProbMed) dataset to rigorously assess LMM performance in medical imaging through probing evaluation and procedural diagnosis. Particularly, probing evaluation features pairing original questions with negation questions with hallucinated attributes, while procedural diagnosis requires reasoning across various diagnostic dimensions for each image, including modality recognition, organ identification, clinical findings, abnormalities, and positional grounding. ProbMed draws from two comprehensive biomedical datasets MedICaT and ChestX-ray14 to compile a diverse set of 6,303 images. These images span three modalities (X-ray, MRI, and CT scan) and four organs (abdomen, brain, chest, and spine). After preprocessing, we generated a diverse set of high-quality questions for each image, covering various diagnostic dimensions. This process resulted in a total of 57,132 question-answer pairs, averaging 9 pairs per image.

Alt text

Dataset Creation

ProbMed was created to rigorously evaluate LMMs’ readiness for real-life diagnostic tasks, particularly under adversarial conditions. Please refer to our huggingface 🤗 Dataset for more details.

Evaluation

Please refer to our eval folder for more details.

🏆 Leaderboard

Model Modality Organ Abnormality Condition/Finding Position Overall
Random Choice 25.00 25.00 50.00 35.67 36.48 32.13
GPT-4o 97.42 69.46 61.79 29.30 24.06 55.60
GPT-4V 92.51 71.73 53.30 35.19 22.40 55.28
Gemini 1.5 Pro 96.47 75.69 62.59 27.93 17.54 55.08
Med-Flamingo 44.15 61.39 50.00 26.33 5.65 35.66
CheXagent 37.25 33.95 73.31 28.52 7.48 30.61
BiomedGPT 60.25 46.81 50.31 14.13 6.11 33.34
LLaVA-Med 5.48 32.96 38.76 20.38 5.33 17.90
MiniGPT-v2 3.25 76.26 50.08 15.23 7.96 27.67
LLaVA-v1.6 (7B) 6.77 80.70 46.18 3.56 1.21 24.96
LLaVA-v1 (7B) 25.27 40.53 50.00 0.34 0.11 19.30

Contact

Citation

BibTeX:

@misc{yan2024worse,
      title={Worse than Random? An Embarrassingly Simple Probing Evaluation of Large Multimodal Models in Medical VQA}, 
      author={Qianqi Yan and Xuehai He and Xiang Yue and Xin Eric Wang},
      year={2024},
      eprint={2405.20421},
      archivePrefix={arXiv},
      primaryClass={cs.AI}
}

Releases

No releases published

Packages

No packages published