Model-Aware Deep Learning for Multimodal and Hyperspectral Imaging

— Professor Aleksandra Pizurica, Dr. Eng. Ghent University, Department of Telecommunications and Information Processing, Group for Artificial Intelligence and Sparse Modelling - GAIM

Abstract
Model-based optimization has long played a central role in inverse imaging and image analysis, providing principled ways to incorporate prior knowledge and domain structure. Today, it increasingly interacts with deep learning: well-founded signal models guide network design, while data-driven training enhances classical inference methods. ...
In this talk, we present learning frameworks that integrate sparse and probabilistic signal models with deep learning for robust image reconstruction and inverse problems in multimodal and hyperspectral imaging. These approaches address compressed sensing and restoration tasks, as well as downstream problems such as classification, clustering, spectral unmixing, and saliency detection within a principled optimization framework. Examples from medical imaging, remote sensing, and art investigation illustrate how model-aware design combined with deep learning improves robustness, interpretability, and practical performance across diverse imaging applications. Read more

Biography
Aleksandra Pižurica is a Full Professor in Statistical Image Modelling at Ghent University, Belgium, where she leads the GAIM group focusing on Artificial Intelligence and Sparse Modelling. She received the Ph.D. degree in Electrical Engineering from Ghent University in 2002. Her research includes multiresolution image models, probabilistic graphical models, sparse coding, representation learning, and image and video reconstruction, restoration, and analysis. ...
Dr. Pižurica was awarded the scientific prize “de Boelpaepe” by the Royal Academy of Science, Letters and Fine Arts of Belgium in 2015 for her contributions to statistical image modelling and applications to digital painting analysis. She was co-recipient of the Best Paper Award of the IEEE GRSS Data Fusion Contest in 2013 and 2014, and the David Hestenes Prize at AGACSE 2018. She is a Senior Area Editor for the IEEE Transactions on Image Processing and previously served as an Associate Editor for the IEEE Transactions on Image Processing and IEEE Transactions on Circuits and Systems for Video Technology. She was the General Chair of the Sixth International Workshop on Image Processing for Art Investigation (IP4AI) in 2018, and has served as TPC Co-Chair of EUSIPCO 2022, Plenary Co-Chair of EUSIPCO 2024 and EUSIPCO 2026, and TPC Co-Chair of ICIP 2026. Read more

Differentiable Digital Twins and AI: Towards the Realization of 6G and Smart Railway Systems

— Dr. techn. Philipp Svoboda Senior Scientist and Head of CDLab AIRAN, Institut for Telecommunication, Technische Universität Wien

Abstract
The transition towards 6G networks necessitates a fundamental paradigm shift from reactive network management towards proactive and autonomous optimization strategies. Central to this evolution is the "Digital Twin" (DT), acting as a high-fidelity virtual representation of the physical radio environment. Utilizing extensive empirical measurements from real-world datasets in Vienna, the construction of differentiable network twins allows for the direct and scalable optimization of critical network parameters—such as transmit power and load-balancing—via gradient-based Artificial Intelligence. ...
​The enhancement of prediction reliability for signal parameters (RSRP), particularly within complex urban environments and railway corridors, is achieved through the integration of uncertainty-aware Bayesian learning. In the context of 6G, these DTs evolve beyond simple monitoring tools to become the core engine for Integrated Sensing and Communication (ISAC), facilitating high-precision localization and context-aware connectivity. By addressing the "sim-to-real" gap as a structured AI challenge, this framework establishes a practical roadmap for sustainable, zero-touch network management. These findings provide a technical basis for future collaboration among researchers focused on digital-twin-based network evolution and intelligent infrastructure. Read more

Biography
Philipp Svoboda is a Researcher at TU Wien, where he serves as the Head of the Christian Doppler Laboratory for Digital Twin assisted Artificial Intelligence for Sustainable Radio Access Networks (CDLab AIRAN). His research aims to contribute to the development of resilient, intelligent, and energy-efficient wireless systems through both fundamental and applied investigations. ...
His current research investigates the integration of Digital Twin (DT) models and Artificial Intelligence (AI) to optimize the design and operation of Radio Access Networks (RANs). This work addresses key challenges identified for 6G and subsequent systems, particularly the enhancement of data-driven network optimization and strategies for improving energy efficiency and environmental sustainability. This focus is informed by his extensive experience analyzing the performance aspects of 4G and 5G cellular technologies. A foundational component of this research involved developing robust frameworks for evaluating network quality using crowdsourcing methodologies, aimed at supporting more reliable connectivity.
Dedicated to the principle of impactful, academic-industrial collaboration, Philipp leads a team of researchers within the CDLab AIRAN. The laboratory actively explores partnerships focused on advancing the state-of-the-art in intelligent and sustainable connectivity and welcomes professional inquiries regarding collaborative research opportunities. Read more