The paper introduces MedNNS, a neural network search framework designed for medical imaging, addressing challenges in architecture selection and weight initialization. MedNNS constructs a meta-space encoding datasets and models based on their performance using a Supernetwork-based approach, expanding the model zoo size by 51x. The framework incorporates rank loss and Fréchet Inception Distance (FID) loss to capture inter-model and inter-dataset relationships, improving alignment in the meta-space and outperforming ImageNet pre-trained DL models and SOTA NAS methods.
MBZUAI researchers are introducing MedNNS, a system to be presented at MICCAI 2025, that recommends the best AI architecture and initialization for medical imaging tasks. MedNNS addresses the challenge of inefficient trial-and-error in building medical imaging AI by reframing model selection as a retrieval problem. The system employs a Once-For-All ResNet-like model and a learned meta-space of 720k model-dataset pairs, using dataset embeddings to predict optimal model performance. Why it matters: By automating model selection, MedNNS promises to significantly reduce the time and resources required to develop effective AI solutions for healthcare, particularly in medical imaging.
A senior lecturer at the University of New South Wales discussed the use of AI to improve early prognosis and personalized treatment plans for neurodegenerative diseases, cardiovascular imaging and multiomics. The lecture highlighted the potential of AI algorithms to detect subtle changes at early stages through advanced multiomics techniques and medical imaging analysis. The speaker has expertise in analyzing medical images and has collaborated with medical professionals to develop AI tools for diagnosis of cancer, neurodegenerative disease, and heart disease. Why it matters: AI-driven prognosis and treatment planning promises earlier intervention and improved outcomes for challenging diseases in the region.
Pierre Baldi from UC Irvine presented applications of AI to biomedicine, covering molecular-level analysis of circadian rhythms, real-time polyp detection in colonoscopy videos, and prediction of post-operative adverse outcomes. He discussed integrating AI in future AI-driven hospitals. The presentation was likely part of a panel discussion hosted by MBZUAI in collaboration with the Manara Center for Coexistence and Dialogue. Why it matters: This highlights the growing interest in AI applications within the healthcare sector in the UAE, particularly through institutions like MBZUAI.
A new brain tumor segmentation method based on convolutional neural networks is proposed for the BraTS-GoAT challenge. The method employs the MedNeXt architecture and model ensembling to segment tumors in brain MRI scans from diverse populations. Experiments on the unseen validation set demonstrate promising results with an average DSC of 85.54%.
Researchers at MBZUAI have introduced MedMerge, a transfer learning technique that merges weights from independently initialized models to improve performance on medical imaging tasks. MedMerge learns kernel-level weights to combine features from different models into a single model. Experiments across various medical imaging tasks demonstrated performance gains of up to 7% in F1 score.
Researchers from MBZUAI have developed MMRINet, a Mamba-based neural network for efficient brain tumor segmentation in MRI scans. The model uses Dual-Path Feature Refinement and Progressive Feature Aggregation to achieve high accuracy with only 2.5M parameters, making it suitable for low-resource clinical environments. MMRINet achieves a Dice score of 0.752 and HD95 of 12.23 on the BraTS-Lighthouse SSA 2025 benchmark.