Saved in:
Bibliographic Details
Main Authors: Nath, Vishwesh, Li, Wenqi, Yang, Dong, Myronenko, Andriy, Zheng, Mingxin, Lu, Yao, Liu, Zhijian, Yin, Hongxu, Tang, Yucheng, Guo, Pengfei, Zhao, Can, Xu, Ziyue, He, Yufan, Heinrich, Greg, Law, Yee Man, Simon, Benjamin, Harmon, Stephanie, Aylward, Stephen, Edgar, Marc, Zephyr, Michael, Han, Song, Molchanov, Pavlo, Turkbey, Baris, Roth, Holger, Xu, Daguang
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2411.12915
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916642758328320
author Nath, Vishwesh
Li, Wenqi
Yang, Dong
Myronenko, Andriy
Zheng, Mingxin
Lu, Yao
Liu, Zhijian
Yin, Hongxu
Tang, Yucheng
Guo, Pengfei
Zhao, Can
Xu, Ziyue
He, Yufan
Heinrich, Greg
Law, Yee Man
Simon, Benjamin
Harmon, Stephanie
Aylward, Stephen
Edgar, Marc
Zephyr, Michael
Han, Song
Molchanov, Pavlo
Turkbey, Baris
Roth, Holger
Xu, Daguang
author_facet Nath, Vishwesh
Li, Wenqi
Yang, Dong
Myronenko, Andriy
Zheng, Mingxin
Lu, Yao
Liu, Zhijian
Yin, Hongxu
Tang, Yucheng
Guo, Pengfei
Zhao, Can
Xu, Ziyue
He, Yufan
Heinrich, Greg
Law, Yee Man
Simon, Benjamin
Harmon, Stephanie
Aylward, Stephen
Edgar, Marc
Zephyr, Michael
Han, Song
Molchanov, Pavlo
Turkbey, Baris
Roth, Holger
Xu, Daguang
contents Generalist vision language models (VLMs) have made significant strides in computer vision, but they fall short in specialized fields like healthcare, where expert knowledge is essential. In traditional computer vision tasks, creative or approximate answers may be acceptable, but in healthcare, precision is paramount.Current large multimodal models like Gemini and GPT-4o are insufficient for medical tasks due to their reliance on memorized internet knowledge rather than the nuanced expertise required in healthcare. VLMs are usually trained in three stages: vision pre-training, vision-language pre-training, and instruction fine-tuning (IFT). IFT has been typically applied using a mixture of generic and healthcare data. In contrast, we propose that for medical VLMs, a fourth stage of specialized IFT is necessary, which focuses on medical data and includes information from domain expert models. Domain expert models developed for medical use are crucial because they are specifically trained for certain clinical tasks, e.g. to detect tumors and classify abnormalities through segmentation and classification, which learn fine-grained features of medical data$-$features that are often too intricate for a VLM to capture effectively especially in radiology. This paper introduces a new framework, VILA-M3, for medical VLMs that utilizes domain knowledge via expert models. Through our experiments, we show an improved state-of-the-art (SOTA) performance with an average improvement of ~9% over the prior SOTA model Med-Gemini and ~6% over models trained on the specific tasks. Our approach emphasizes the importance of domain expertise in creating precise, reliable VLMs for medical applications.
format Preprint
id arxiv_https___arxiv_org_abs_2411_12915
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle VILA-M3: Enhancing Vision-Language Models with Medical Expert Knowledge
Nath, Vishwesh
Li, Wenqi
Yang, Dong
Myronenko, Andriy
Zheng, Mingxin
Lu, Yao
Liu, Zhijian
Yin, Hongxu
Tang, Yucheng
Guo, Pengfei
Zhao, Can
Xu, Ziyue
He, Yufan
Heinrich, Greg
Law, Yee Man
Simon, Benjamin
Harmon, Stephanie
Aylward, Stephen
Edgar, Marc
Zephyr, Michael
Han, Song
Molchanov, Pavlo
Turkbey, Baris
Roth, Holger
Xu, Daguang
Computer Vision and Pattern Recognition
Generalist vision language models (VLMs) have made significant strides in computer vision, but they fall short in specialized fields like healthcare, where expert knowledge is essential. In traditional computer vision tasks, creative or approximate answers may be acceptable, but in healthcare, precision is paramount.Current large multimodal models like Gemini and GPT-4o are insufficient for medical tasks due to their reliance on memorized internet knowledge rather than the nuanced expertise required in healthcare. VLMs are usually trained in three stages: vision pre-training, vision-language pre-training, and instruction fine-tuning (IFT). IFT has been typically applied using a mixture of generic and healthcare data. In contrast, we propose that for medical VLMs, a fourth stage of specialized IFT is necessary, which focuses on medical data and includes information from domain expert models. Domain expert models developed for medical use are crucial because they are specifically trained for certain clinical tasks, e.g. to detect tumors and classify abnormalities through segmentation and classification, which learn fine-grained features of medical data$-$features that are often too intricate for a VLM to capture effectively especially in radiology. This paper introduces a new framework, VILA-M3, for medical VLMs that utilizes domain knowledge via expert models. Through our experiments, we show an improved state-of-the-art (SOTA) performance with an average improvement of ~9% over the prior SOTA model Med-Gemini and ~6% over models trained on the specific tasks. Our approach emphasizes the importance of domain expertise in creating precise, reliable VLMs for medical applications.
title VILA-M3: Enhancing Vision-Language Models with Medical Expert Knowledge
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2411.12915