Resilient Artificial Intelligence in Medical Imaging: Handling and Mining Multiple Sources

Title

Resilient Artificial Intelligence in Medical Imaging: Handling and Mining Multiple Sources

Abstract

In the recent years, Artificial Intelligence (AI), and in particular Machine Learning (ML) and Deep Learning (DL), had a radical spread in medical image computing with surprising results. Moreover, the use of deep neural networks has also enabled the development of DL-based solutions in medical applications characterized by the need of leveraging information coming from multimodal data sources, raising the Multimodal Deep Learning (MDL). Resilience in multimodal artificial intelligence (AI) is essential for healthcare, where data is often incomplete, variable, and diverse. In real-world clinical settings, one or more data modalities—such as imaging, lab results, or wearable data—may be missing or inconsistent due to resource constraints, patient factors, or technological limitations. Resilient AI systems can adapt to these challenges by using data imputation, synthetic generation, and flexible model architectures that maintain performance despite gaps or variability. Furthermore, resilience allows these systems to address and reduce biases, ensuring that insights are fair and applicable across all patient demographics. By achieving this robustness, resilient multimodal AI not only supports consistent, accurate diagnostics but also helps build trust in AI-driven healthcare, making it scalable and reliable in diverse clinical environments.

Bio

Michela Gravina is an Assistant Professor (non-tenure track RTD-A) at the University of Naples Federico II, Department of Electrical Engineering and Information Technology. She earned her Ph.D. in Information and Communication Technology for Health (ICTH) at the University of Naples Federico II, focusing on artificial intelligence applications in medical imaging, particularly in handling multiple and heterogeneous data sources. She graduated with honors in Computer Science. Her primary research interests include image analysis and recognition, machine learning, and deep learning. Her research in medical image computing focuses on multimodal deep learning, physiologically-aware synthetic data generation, and resilient artificial intelligence.

When

Monday November 18th, 12:00

Where

Room 322, UniGe DIBRIS/DIMA, Via Dodecaneso 35