Συντάχθηκε 08-05-2026 14:21
Τόπος: Λ - Κτίριο Επιστημών/ΗΜΜΥ, 141Π-98
Έναρξη: 11/05/2026 14:00
Λήξη: 11/05/2026 15:00
Abstract
The performance and efficiency of modern signal processing systems depend on tight and often non-linear interactions between algorithms and hardware. Algorithmic gains that look compelling on paper can diminish or even reverse when confronted with real architectural constraints, just as new hardware may underperform without algorithms designed to exploit it. Evaluations that focus on only one dimension therefore risk producing misleading conclusions about scalability, energy efficiency, or impact. This talk argues that algorithm–hardware co-design is essential not only for optimization, but for reaching correct insights. We will focus on three signal processing applications (decoding of polar codes, non-linearity compensation, and decoding of Reed-Muller codes), each showcasing a different aspect where the joint consideration of algorithms and hardware can produce unexpected results.
About the Speaker
Dr. Alexios Balatsoukas-Stimming holds Dipl.-Ing. (2010) and MSc (2012) degrees in Electronics and Computer Engineering from the Technical University of Crete. He received his PhD in Computer and Communications Sciences from the École polytechnique fédérale de Lausanne (EPFL) in 2016. He then spent a year as Marie Skłodowska-Curie postdoctoral fellow at the European Laboratory for Particle Physics (CERN), before joining the Telecommunications Circuits Laboratory of EPFL as postdoctoral researcher. Dr. Balatsoukas-Stimming is currently a tenured Assistant Professor in the Electronic Systems Group of the Eindhoven University of Technology. His research lies on the intersection between communications, hardware design, and machine learning and his specific research interests include VLSI circuits for communications, error-correction coding theory and practice, and non-linear signal processing.
Σύνδεσμος εκδήλωσης
