Neuromorphic computing and signal processing training network

The exponential surge in artificial intelligence (AI), internet traffic, and online services demands a revolutionary leap in devices, computing architecture, and integration technologies. Current digital computing platforms, driving societal progress, are hitting fundamental limits such as quantum tunneling and facing technical challenges like energy efficiency and heat dissipation. As a result, the progress in AI is slowed down by speed bottlenecks, massive energy consumption, and limited access to AI technologies as only selected large industrial players can afford to commit the required resources. Inspired by the brain’s powerful and energy-efficient processing capabilities, neuromorphic computing presents a promising solution to these critical issues. However, to fully harness neuromorphic computing’s potential, a holistic optimization - from individual computing devices to the overall architecture, including a focus on applications, and training methods - across multiple technological platforms - photonics, electronics, biological neurons - is critically needed. The proposal of MINDnet addresses these urgent challenges through an interdisciplinary approach, bridging traditionally separate fields. MINDnet’s mission is to train 15 Doctoral Candidates, forming the next generation of leading scientists and expert researchers in the field of neuromorphic computing. MINDnet leans on a consortium of 10 beneficiaries - 7 universities, 1 tech-transfer research center, 1 multinational, 1 SME - complemented by 6 associated partners – 1 multinational, 2 SMEs, 2 tech-transfer research centers, and 1 academic. The consortium spans over 8 European countries, combining experts in 5 research disciplines - photonics (DTU, UNITN, USTRATH), electronics (UCL, HPE, FZJ, SCS), nonlinear dynamics (TUIL) and machine learning (UNIPI, TUG), - as well as a strong focus on the 5 applications - telecom (NVIDIA, HHI), sensing (FCAP), geolocation (AT), space (NB), and biomedical (UNITN).