From December 9 to 11, 2025, the research team from the Faculty of Electronic Engineering, University of Niš (FEEUNI), visited the IHP – Leibniz Institute for High Performance Microelectronics in Frankfurt (Oder), Germany, as part of the AIDA4Edge project.
We were warmly welcomed by Marko Anđelković on behalf of IHP. The visit began with an overview of the AIDA4Edge project, presented by project coordinator Prof. Milan Dinčić, followed by a series of technical presentations by IHP researchers on key topics in AI hardware acceleration and system design. Jianan Wen presented research on energy-efficient and reliable RRAM-based in-memory computing architectures for AI acceleration. Alessandro Veronesi discussed the design of deep learning systems using modern hardware description frameworks, while Rizwan Tariq Syed presented a reconfigurable system architecture aimed at flexible and efficient AI processing.
The visit also included a two-way exchange of knowledge. On the second day, members of the FEEUNI research team presented their recent research results and ongoing activities. Aleksandra Jovanović introduced a concept for real-to-integer data conversion using uniform quantization tailored to the statistical characteristics of data in neural network MAC computations. Sofija Perić presented results on the application of power-of-two quantizers and a 3-bit integer quantizer for weight quantization, evaluating their impact on accuracies of spiking and artificial neural networks. Jelena Nedeljković provided an overview of the current research status within the joint mentorship framework between FEEUNI and IHP, highlighting progress and future research directions.
The agenda also included dedicated meetings on joint research activities, ESR mentorship, and preparation of future project proposals. These discussions enabled alignment of ongoing efforts, identification of new collaboration opportunities, and definition of next steps within and beyond the AIDA4Edge project.
Overall, the visit facilitated a structured exchange of expertise, strengthened the partnership between FEEUNI and IHP, and laid the groundwork for deeper collaboration in future research initiatives.
We extend our gratitude to Marko Anđelković and the IHP team for sharing their expertise and fostering a welcoming environment throughout the visit.

