AIDA4Edge Contributions at SSSS 2026: Resilient IC Design and Efficient GNN Quantization

We are pleased to share that the Small Systems Simulation Symposium (SSSS 2026) was successfully held from February 12–14, 2026, at the Faculty of Electronic Engineering, University of Niš.

Within the framework of the AIDA4Edge project, our research team contributed with two significant presentations addressing current challenges in quantization and resilient hardware systems.

  • Invited Keynote Presentation
    Marko Andjelković, Rizwan Tariq Syed, Markus Ulbricht, Fabian Vargas, Zoran Perić, Milan Dincić, Jelena Nikolić, Aleksandra Jovanović, Dejan Cirić, Tatjana Nikolić, Goran Nikolić, Jelena Nedeljković, Sofija Perić, Nikola Vučić, Miloš Krstić
    “An Overview of On-Chip Sensing and Analytics for Resilient Integrated Circuits”

The invited keynote was delivered by Marko Andjelković, providing a comprehensive overview of on-chip sensing methodologies and embedded analytics aimed at improving the resilience, reliability, and adaptability of modern integrated circuits.

  • Paper Presentation
    Zoran Perić, Nikola Simić, Milan Tančić, Sofija Perić, Oliver Rhodes, Ivan Popović
    “Analysis of Classification Accuracy of FxP 8-bit and FP 8-bit Quantizers in GNNs”

The paper was presented by Sofija Perić, focusing on a comparative evaluation of fixed-point (FxP) and floating-point (FP) 8-bit quantization strategies in Graph Neural Networks (GNNs). The study provides insight into classification accuracy trade-offs and practical implications for efficient edge-AI deployment.

Participation in SSSS 2026 further strengthened interdisciplinary collaboration within AIDA4Edge and contributed to ongoing research efforts in quantization-aware AI systems and resilient electronic design.

We thank the organizers and participants for fostering an engaging and high-quality scientific exchange.