On April 15, 2026, the Faculty of Electronic Engineering, University of Niš hosted the 2nd AIDA4Edge Project Workshop entitled “Efficient and Reliable Edge AI: From Quantization and Optimization to Neuromorphic Computing.” The workshop was held in a hybrid format, both in person and online, and gathered 32 participants (25 onsite and 7 online), including project partners, researchers, and a significant number of stakeholders. The event served as an important platform for the dissemination of project results and ongoing research activities within AIDA4Edge.
The workshop was opened by Prof. Milan Dinčić, AIDA4Edge project coordinator, who also delivered a presentation on fixed-point format optimization, addressing efficient numerical representations for neural networks. In continuation, Prof. Riccardo Zese presented approaches to optimizing neural networks for edge devices, focusing on balancing performance and resource constraints.
The program then moved towards emerging computing paradigms, where Dr. Edward Jones introduced recent advances in event-based sensing and spiking neural networks, highlighting low-latency and energy-efficient vision paradigms. Complementing this hardware-oriented perspective, Dr. Rizwan Tariq Syed presented FPGA-based workflows for DNN acceleration, demonstrating the path from high-level models to hardware implementation.
Further extending the discussion on efficiency, subsequent sessions addressed advanced topics such as dynamic quantization of binary neural networks for pixel processor arrays, presented by Yizhen Bi, and network efficiency through pruning, presented by Prof. Stefan Tomić together with Prof. Jelena Nikolić, emphasizing complementary strategies for reducing computational complexity while maintaining performance.
The participation of project stakeholders, including Dr. Miloš Marković (ProMetronics GmbH), Prof. Nikola Simić (Faculty of Technical Sciences, University of Novi Sad), Prof. Stefan Tomić (Abu Dhabi Polytechnic), Dr. Milan Grujev (ARTIFACT project), Prof. Marko Petković (Faculty of Science and Mathematics, University of Niš), Đorđe Antić (City of Niš), and Milena Pecarski (Mpeeria), further enriched the discussions and strengthened the connection between research and application.
The workshop concluded with a dedicated Early Stage Researchers (ESRs) session, featuring presentations by Sofija Perić, Jelena Nedeljković, Tamara Milovanović, and Nikola Vučić. During this session, ESRs presented their current results, challenges, and future research directions, and received valuable feedback and guidance from senior researchers within the project, which will significantly support their further work and development.
Overall, the workshop provided a valuable opportunity for the dissemination of results, exchange of ideas, and strengthening of collaboration among partners and stakeholders, contributing to the advancement of efficient, reliable, and hardware-aware Edge AI solutions.
