Jun
19
Date: 19-June-2026
Time: 10:00 AM ET (New York Time)
Presenter: Dr. Nikos Deligiannis
Website link: Register
Registration link:
https://tudelft.zoom.us/meeting/register/tJwvceuqqz4tH9AH4UPShaMbXp9qmW4pCy3x
The ASI Webinar Series is an event initiated by the Autonomous System Initiative (ASI) of the IEEE Signal Processing (SP) Society. The goal is to offer the SP community with free webinars looking into the future of autonomous systems. These monthly webinars are hosted on Zoom, with recordings made available in the IEEE ASI’s YouTube channelfollowing the live events.
Abstract
What Signal Processing Brings to Efficient and Explainable AI for Autonomous Systems
Autonomous systems, such as autonomous ground and aerial vehicles, increasingly rely on large-scale AI models to perceive, communicate, and act in complex environments. Yet, these systems must operate under strict constraints: limited bandwidth, limited energy, uncertain sensing conditions, and the need for transparent and reliable decisions. In this talk, the presenter will discuss how signal processing can provide the missing structure needed to make AI for autonomous systems more efficient, interpretable, and deployable.
The presenter will present a sequence of recent works from their group that combine model-based signal processing with modern deep learning. First, he will discuss neural rate-adaptive Slepian–Wolf decoding, in which Transformer models decode LDPC-based syndrome codes for distributed source coding, enabling efficient compression of correlated camera streams in distributed perception. Then move to deep unfolding Transformers for sparse video recovery, showing how attention-based architectures can be interpreted as trainable sparse-recovery algorithms for video reconstruction and denoising. Next, he will present unfolded RPCA networks with foreground masking for interpretable background subtraction and foreground detection, with applications in traffic monitoring and robotics, then discuss how these model-based architectures can be converted into power-efficient spiking implementations for edge deployment.
The talk will conclude by broadening the discussion to explainable and semantic perception, including model-agnostic visual explanations based on approximate bilinear models and zero-shot referring multi-object tracking with vision-language models. Overall, the talk argues that signal processing remains central to autonomous systems: it offers principled ways to encode domain knowledge, reduce communication and computation, improve interpretability, and bridge low-level sensing with high-level semantic understanding.
Biography
Nikos Deligiannis (SM’25) received the diploma degree in electrical and computer engineering from the University of Patras, Patras, Greece, in 2006, and the Ph.D. degree in engineering sciences from Vrije Universiteit Brussel (VUB), Brussels, Belgium, in 2012. From 2013 to 2015, he was a Postdoctoral Researcher in the Department of Electronic and Electrical Engineering at University College London, London, U.K.
He is currently a Professor with the Department of Electronics and Informatics (ETRO), Vrije Universiteit Brussel (VUB), the holder of the 2024-2025 Francqui Research Professorship on Trustworthy AI at VUB, and a Principal Investigator (PI) at imec, Belgium. He is the PI of the ERC Consolidator Grant IONIAN, conducting research at the intersection between interpretable and explainable AI and multiterminal compression for autonomous vehicles.
Dr. Deligiannis is a senior member of the IEEE and member of EURASIP and served as the Chair of the EURASIP Technical Area Committee on Signal and Data Analytics for Machine Learning in 2021-2023. He serves as an Associate Editor for the IEEE Transactions on Image Processing and has been a Guest Editor for special issues at the EURASIP Journal on Advances in Signal Processing and the Signal Processing journal.
