NimbleAI

Thu, 01 Dec 2022 00:00:00 +0000

🌐 Project Webpage: | 🇪🇺 CORDIS Factsheet:

NimbleAI (Ultra-energy efficient and secure neuromorphic sensing and processing at the endpoint) is a Horizon Europe research project (Grant Agreement ID 101070679) funded under the HORIZON-CL4-2021-DIGITAL-EMERGING-01 programme.

📅 Project Timeline & Factsheet

Start Date: October 1, 2022
End Date: March 31, 2026
Duration: 42 Months
Funding Scheme: Horizon Europe (Research and Innovation Action - HORIZON-RIA)
Grant Agreement ID: 101070679
Overall Budget: €6,000,000+ (EU Contribution)

🎯 Objectives & Scope

Traditional visual computing pipelines (cameras connected to separate CPU/GPUs) are highly inefficient, wasting substantial energy transmitting redundant video frames. NimbleAI developed an energy-efficient sensing-processing system operating on event-driven, bio-inspired principles.

💡 Key Research Innovations

Event-Based Sensing: Integrated dynamic event vision sensors that process visual changes at pixel level, eliminating redundant frame capture.
3D Neuromorphic Integration: Stacked sensing layers directly above event-driven in-memory neural processing arrays.
Adaptive Edge AI: Enabled dynamic trade-offs between energy consumption, latency, and inference accuracy for autonomous robotics and IoT nodes.

🏛️ BSC Key Role & Research Deliverables

As a core partner in NimbleAI, Barcelona Supercomputing Center (BSC) led critical research thrusts in neuromorphic hardware architecture and edge AI dependability:

Neuromorphic Execution Trade-Offs: Co-designed the architectural framework balancing bio-inspired event-driven vision processing with 3D stacked in-memory neural acceleration.
Predictability & Dependability: Developed safety and execution models to guarantee bounded response times for edge AI inferences.
Flagship Publication: Led and co-authored the flagship project paper at Design, Automation and Test in Europe (DATE 2023): “NimbleAI: Towards Neuromorphic Sensing-Processing 3D-integrated Chips”.

📚 Associated Publications

X. Iturbe, N. Abderrahmane, J. Abella, S. Alcaide, et al. — DATE 2023 ( )

🏛️ Consortium Partners

(Partners are listed in no particular order)

The NimbleAI consortium brings together 19 leading European research centers, universities, semiconductor companies, and industrial leaders extracted directly from the official project website ( ):

🇪🇸 (Spain, Coordinator)
🇪🇸 (Spain)
🇪🇸 (Spain)
🇪🇸 (Spain)
🇪🇸 (Spain)
🇬🇧 (United Kingdom)
🇬🇧 (United Kingdom)
🇫🇷 (France)
🇫🇷 (France)
🇧🇪 (Belgium)
🇦🇹 (Austria)
🇦🇹 (Austria)
🇦🇹 (Austria)
🇳🇱 (Netherlands)
🇳🇱 (Netherlands)
🇩🇪 (Germany / Czech Republic)
🇩🇪 (Germany)
🇮🇹 (Italy)
🇮🇹 (Italy)
🇺🇸 (United States / Austria)

💼 Industry Advisory Board & Partners

🇪🇺
🇳🇱
🇺🇸
🇺🇸
🇫🇷
🇺🇸

Edge AI |