The field of multimodal learning is rapidly advancing, with a focus on medical applications. Recent research has explored the use of large language models, vision-language models, and multimodal fusion techniques to improve performance on various medical tasks, such as disease diagnosis, image segmentation, and question answering. Notably, the development of specialized models like Med-GRIM, VL-MedGuide, and Doctor Sun has demonstrated significant improvements in medical visual question answering and image classification tasks. Furthermore, the introduction of datasets like Med-GLIP-5M and MM-Food-100K has facilitated the training and evaluation of multimodal models for medical applications. Overall, the field is moving towards more effective and interpretable multimodal learning approaches for medical decision support. Noteworthy papers include Med-GRIM, which achieved state-of-the-art performance on medical VQA tasks, and VL-MedGuide, which demonstrated strong performance on skin disease diagnosis and concept detection.
Advances in Multimodal Learning for Medical Applications
Sources
PEACH: A sentence-aligned Parallel English-Arabic Corpus for Healthcare
SIFThinker: Spatially-Aware Image Focus for Visual Reasoning
CLIPin: A Non-contrastive Plug-in to CLIP for Multimodal Semantic Alignment
Harnessing Adaptive Topology Representations for Zero-Shot Graph Question Answering
Med-GRIM: Enhanced Zero-Shot Medical VQA using prompt-embedded Multimodal Graph RAG
Large Language Models Facilitate Vision Reflection in Image Classification
On the effectiveness of multimodal privileged knowledge distillation in two vision transformer based diagnostic applications
VL-MedGuide: A Visual-Linguistic Large Model for Intelligent and Explainable Skin Disease Auxiliary Diagnosis
BASIC: Boosting Visual Alignment with Intrinsic Refined Embeddings in Multimodal Large Language Models
MV-CoRe: Multimodal Visual-Conceptual Reasoning for Complex Visual Question Answering
MMReID-Bench: Unleashing the Power of MLLMs for Effective and Versatile Person Re-identification
Low-Rank Expert Merging for Multi-Source Domain Adaptation in Person Re-Identification
FLUID: Flow-Latent Unified Integration via Token Distillation for Expert Specialization in Multimodal Learning
MIMIC: Multimodal Inversion for Model Interpretation and Conceptualization
Investigating the Design Space of Visual Grounding in Multimodal Large Language Model
The Medical Metaphors Corpus (MCC)
Information Bottleneck-based Causal Attention for Multi-label Medical Image Recognition
MedReasoner: Reinforcement Learning Drives Reasoning Grounding from Clinical Thought to Pixel-Level Precision
Capabilities of GPT-5 on Multimodal Medical Reasoning
Spatial-ORMLLM: Improve Spatial Relation Understanding in the Operating Room with Multimodal Large Language Model
Doctor Sun: A Bilingual Multimodal Large Language Model for Biomedical AI
AME: Aligned Manifold Entropy for Robust Vision-Language Distillation
MMIF-AMIN: Adaptive Loss-Driven Multi-Scale Invertible Dense Network for Multimodal Medical Image Fusion
GazeLT: Visual attention-guided long-tailed disease classification in chest radiographs
Multi-Contrast Fusion Module: An attention mechanism integrating multi-contrast features for fetal torso plane classification
Performance of GPT-5 Frontier Models in Ophthalmology Question Answering
Interpretable Oracle Bone Script Decipherment through Radical and Pictographic Analysis with LVLMs
MANGO: Multimodal Attention-based Normalizing Flow Approach to Fusion Learning
MM-Food-100K: A 100,000-Sample Multimodal Food Intelligence Dataset with Verifiable Provenance
Med-GLIP: Advancing Medical Language-Image Pre-training with Large-scale Grounded Dataset
Performance of GPT-5 in Brain Tumor MRI Reasoning
Medico 2025: Visual Question Answering for Gastrointestinal Imaging