On the Compositional Generalization of Multimodal LLMs for Medical Imaging

• URL: http://arxiv.org/abs/2412.20070v1
• Date: Sat, 28 Dec 2024 07:50:00 GMT
• Title: On the Compositional Generalization of Multimodal LLMs for Medical Imaging
• Authors: Zhenyang Cai, Junying Chen, Rongsheng Wang, Weihong Wang, Yonglin Deng, Dingjie Song, Yize Chen, Zixu Zhang, Benyou Wang,
• Abstract summary: Multimodal large language models (MLLMs) hold significant potential in the medical field.<n>Their capabilities are often limited by insufficient data in certain medical domains, highlighting the need for understanding what kinds of images can be used by MLLMs for generalization.<n>This paper attempts to employ compositional generalization (CG)-the ability of models to understand novel combinations by recombining learned elements-as a guiding framework.<n>Experiments confirmed that MLLMs can use CG to understand unseen medical images and identified CG as one of the main drivers of the generalization observed in multi-task training.
• Score: 14.419190976672065
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Multimodal large language models (MLLMs) hold significant potential in the medical field, but their capabilities are often limited by insufficient data in certain medical domains, highlighting the need for understanding what kinds of images can be used by MLLMs for generalization. Current research suggests that multi-task training outperforms single-task as different tasks can benefit each other, but they often overlook the internal relationships within these tasks, providing limited guidance on selecting datasets to enhance specific tasks. To analyze this phenomenon, we attempted to employ compositional generalization (CG)-the ability of models to understand novel combinations by recombining learned elements-as a guiding framework. Since medical images can be precisely defined by Modality, Anatomical area, and Task, naturally providing an environment for exploring CG. Therefore, we assembled 106 medical datasets to create Med-MAT for comprehensive experiments. The experiments confirmed that MLLMs can use CG to understand unseen medical images and identified CG as one of the main drivers of the generalization observed in multi-task training. Additionally, further studies demonstrated that CG effectively supports datasets with limited data and delivers consistent performance across different backbones, highlighting its versatility and broad applicability. Med-MAT is publicly available at https://github.com/FreedomIntelligence/Med-MAT.

Related papers

• Zeus: Zero-shot LLM Instruction for Union Segmentation in Multimodal Medical Imaging [4.341503087761129]
  Conducting multimodal learning involves visual and text modalities shown as a solution, but collecting paired vision-language datasets is expensive and time-consuming. Inspired by the superior ability in numerous cross-modal tasks for Large Language Models (LLMs), we proposed a novel Vision-LLM union framework to address the issues.
  arXiv  Detail & Related papers  (2025-04-09T23:33:35Z)
• UMIT: Unifying Medical Imaging Tasks via Vision-Language Models [17.65946656129399]
  UMIT is a unified multi-modal, multi-task VLM designed specifically for medical imaging tasks. It is able to solve various tasks, including visual question answering, disease detection, and medical report generation. It supports both English and Chinese, expanding its applicability globally.
  arXiv  Detail & Related papers  (2025-03-20T06:43:36Z)
• MRGen: Segmentation Data Engine For Underrepresented MRI Modalities [59.61465292965639]
  Training medical image segmentation models for rare yet clinically significant imaging modalities is challenging due to the scarcity of annotated data. This paper investigates leveraging generative models to synthesize training data, to train segmentation models for underrepresented modalities.
  arXiv  Detail & Related papers  (2024-12-04T16:34:22Z)
• A Survey of Medical Vision-and-Language Applications and Their Techniques [48.268198631277315]
  Medical vision-and-language models (MVLMs) have attracted substantial interest due to their capability to offer a natural language interface for interpreting complex medical data. Here, we provide a comprehensive overview of MVLMs and the various medical tasks to which they have been applied. We also examine the datasets used for these tasks and compare the performance of different models based on standardized evaluation metrics.
  arXiv  Detail & Related papers  (2024-11-19T03:27:05Z)
• MOSMOS: Multi-organ segmentation facilitated by medical report supervision [10.396987980136602]
  We propose a novel pre-training & fine-tuning framework for Multi-Organ Supervision (MOS) Specifically, we first introduce global contrastive learning to align medical image-report pairs in the pre-training stage. To remedy the discrepancy, we further leverage multi-label recognition to implicitly learn the semantic correspondence between image pixels and organ tags.
  arXiv  Detail & Related papers  (2024-09-04T03:46:17Z)
• MMSci: A Dataset for Graduate-Level Multi-Discipline Multimodal Scientific Understanding [59.41495657570397]
  This dataset includes figures such as schematic diagrams, simulated images, macroscopic/microscopic photos, and experimental visualizations. We developed benchmarks for scientific figure captioning and multiple-choice questions, evaluating six proprietary and over ten open-source models. The dataset and benchmarks will be released to support further research.
  arXiv  Detail & Related papers  (2024-07-06T00:40:53Z)
• Med-MoE: Mixture of Domain-Specific Experts for Lightweight Medical Vision-Language Models [17.643421997037514]
  We propose a novel framework that tackles both discriminative and generative multimodal medical tasks. The learning of Med-MoE consists of three steps: multimodal medical alignment, instruction tuning and routing, and domain-specific MoE tuning. Our model can achieve performance superior to or on par with state-of-the-art baselines.
  arXiv  Detail & Related papers  (2024-04-16T02:35:17Z)
• Eye-gaze Guided Multi-modal Alignment for Medical Representation Learning [65.54680361074882]
  Eye-gaze Guided Multi-modal Alignment (EGMA) framework harnesses eye-gaze data for better alignment of medical visual and textual features. We conduct downstream tasks of image classification and image-text retrieval on four medical datasets.
  arXiv  Detail & Related papers  (2024-03-19T03:59:14Z)
• Enhancing Representation in Radiography-Reports Foundation Model: A Granular Alignment Algorithm Using Masked Contrastive Learning [26.425784890859738]
  MaCo is a masked contrastive chest X-ray foundation model. It simultaneously achieves fine-grained image understanding and zero-shot learning for a variety of medical imaging tasks. It is shown to be superior over 10 state-of-the-art approaches across tasks such as classification, segmentation, detection, and phrase grounding.
  arXiv  Detail & Related papers  (2023-09-12T01:29:37Z)
• LVM-Med: Learning Large-Scale Self-Supervised Vision Models for Medical Imaging via Second-order Graph Matching [59.01894976615714]
  We introduce LVM-Med, the first family of deep networks trained on large-scale medical datasets. We have collected approximately 1.3 million medical images from 55 publicly available datasets. LVM-Med empirically outperforms a number of state-of-the-art supervised, self-supervised, and foundation models.
  arXiv  Detail & Related papers  (2023-06-20T22:21:34Z)
• Modality-Agnostic Learning for Medical Image Segmentation Using Multi-modality Self-distillation [1.815047691981538]
  We propose a novel framework, Modality-Agnostic learning through Multi-modality Self-dist-illation (MAG-MS) MAG-MS distills knowledge from the fusion of multiple modalities and applies it to enhance representation learning for individual modalities. Our experiments on benchmark datasets demonstrate the high efficiency of MAG-MS and its superior segmentation performance.
  arXiv  Detail & Related papers  (2023-06-06T14:48:50Z)
• Towards Medical Artificial General Intelligence via Knowledge-Enhanced Multimodal Pretraining [121.89793208683625]
  Medical artificial general intelligence (MAGI) enables one foundation model to solve different medical tasks. We propose a new paradigm called Medical-knedge-enhanced mulTimOdal pretRaining (MOTOR)
  arXiv  Detail & Related papers  (2023-04-26T01:26:19Z)
• Cross-Modal Information Maximization for Medical Imaging: CMIM [62.28852442561818]
  In hospitals, data are siloed to specific information systems that make the same information available under different modalities. This offers unique opportunities to obtain and use at train-time those multiple views of the same information that might not always be available at test-time. We propose an innovative framework that makes the most of available data by learning good representations of a multi-modal input that are resilient to modality dropping at test-time.
  arXiv  Detail & Related papers  (2020-10-20T20:05:35Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.