NeuroSys – Neuromorphic hardware for autonomous artificial intelligence systems
Prof. Dr.-Ing. Max Lemme
- Send Email
NeuroSys is part of the future cluster initiative “Clusters4Future” of the German Federal Ministry of Education and Research (BMBF). The initiative is composed of three three-year implementation phases, with the first implementation phase started on 01.01.2022.
NeuroSys has the vision to establish the Aachen region as a world’s leading location for research, development, and innovation in neuromorphic hardware for artificial intelligence (AI). For this purpose, all competencies are bundled in the region and built up into an innovation ecosystem needed for the development of future European AI hardware. The long-term vision is to unfold a European path through technological independence in this ethically and economically sensitive area. AI, implemented as software, already dominates areas such as computer vision and speech processing. However, innovative new hardware concepts are needed to sustainably realize applications such as autonomous driving, personalized healthcare, smart cities, the Internet of Things, and Economy 4.0, because conventional computer hardware is increasingly running up against inherent limits in energy efficiency for AI applications. NeuroSys overcomes these limits by developing neuro-inspired hardware that can revolutionize AI systems in terms of energy efficiency and performance. NeuroSys encompasses a broad spectrum of expertise that can initiate an integrated and sustainable research and transformation process through interdisciplinary interlocking. Physicists, material- and neuroscientists, engineers and computer scientists work together with economists, ethicists and sociologists on innovations that are not only technologically advanced, but at the same time economically value-adding as well as socially useful and desirable. RWTH Aachen University, as the nucleus, works closely with the Helmholtz Institute, the FZ-Jülich and the NRW Landesinstitut AMO GmbH. Start-ups and regional companies are an integral part of the cluster, while nationwide corporations and internationally renowned scientists make up the advisory board.
In detail, NeuroSys consists of five subprojects (A - E) with different focal points that work closely together:
Project A: Memristor Crossbar Architectures
- In Project A, the technology for fabricating memristive crossbar architectures is being developed. This includes the optimization of different types of memristive materials and devices (oxides and 2D materials) and the development of schemes for their hybrid integration into CMOS back-end-of-line (BEOL) processes.
Project B: Photonic Neuromorphic Circuits.
- Optical transmission systems enable extremely high data rates and substantial reductions in latency during signal transmission. Here, these properties are used to accelerate neuromorphic hardware far beyond the performance of pure electronics. In parallel, work is underway to integrate photonics into the BEOL process and to develop remanent optical control elements for later combination with the system concept.
Project C: Algorithm-Hardware Co-Design
- Project C brings together expertise from neuroscience, automated system design, and hardware-related circuit design to apply neuromorphic principles to the development of novel algorithms and novel device characteristics.
Project D: Impulses Through Applications.
- The goal of Project D is to investigate/analyze the benefits of neuromorphic hardware for optimal use from the user perspective (application-pull). To this end, four representative application examples of neuromorphic hardware for different requirements will be explored:
- Camera-based diagnostics
- Speech translation
- Video object segmentation
- Individualized medical controllers
Project E: Innovation Process and Business Model Development
- In Project E, the development of technical architectures, circuits, and applications is systematically extended to include economic, societal, and ethical analyses. These considerations are essential to ensure an economically viable, ethically desirable, and socially feasible development process for the innovations emerging from Projects A-D.
The overall coordination of the cluster runs through our chair. Therefore, if you have any questions, please feel free to contact us directly at firstname.lastname@example.org. Further information can be found on the project website https://www.neurosys.info.
In addition, our chair is involved in projects A and E. In project A we develop fabrication and characterization processes for memristive arrays based on 2D materials and their integration into CMOS structures. In addition, we develop a fabrication process for crossbar arrays in the so-called 1-transistor-1-memristor structure with memristive devices in the nanometer range. Within project E, our chair is the interface between the technical-scientific actors on the one hand and the economic and socio-scientific as well as innovation-accompanying actors on the other hand. This ensures the fast and smooth transfer of knowledge from NeuroSys research to the economic and social scientists of the initiative.
The NeuroSys future cluster is funded by the German Federal Ministry of Education and Research (BMBF).
RWTH Aachen University, Aachen, Germany
- Compound Semiconductor Technology (CST)
- Institute of Integrated Photonics (IPH)
- I. Institute of Physics (IA)
- Chair of Integrated Digital Systems and Circuit Design (IDS)
- The Chair of Integrated Analog Circuits and RF Systems (IAS)
- Chair of High Frequency Electronics (HFE)
- Chair for Software for Systems on Silicon (SSS)
- Medical Information Technology (MedIT)
- Chair of Computer Science 6: Human Language Technology and Pattern Recognition (HLTPR)
- Chair of Computer Science 13: Computer Vision (CVG)
- Chair of Information Theory and Data Analytics (INDA)
- Chair of Management Accounting (Controlling)
- Chair of Society and Technology (SoTec)
- LuF Applied Ethics (Ethics)
Forschungszentrum Jülich, Jülich, Germany
- Electronic Materials (PGI-7)
- JARA-Institute Energy-efficient information technology (Green IT) (PGI-10)
- Institute of Neuroscience and Medicine (INM-6)
- Neuromorphic Compute Nodes (PGI-14)
- Neuromorphic Software Ecosystems (PGI-15)
- AMO GmbH, Aachen, Germany
- AiXscale Photonics UG, Köln, Germany
- Black Semiconductor GmbH, Aachen, Germany
- Clinomic Medical GmbH, Aachen, Germany
- AixACCT Systems GmbH, Aachen, Germany
- AppTek GmbH, Aachen, Germany
- Gremse IT, Aachen, Germany
- RWTH Innovation GmbH, Aachen, Germany
- STAR Healthcare Management, Köln, Germany
- AIXTRON SE, Herzogenrath, Germany
- ELMOS Semiconductor SE, Dortmund, Germany (associated partner)
- IHK Aachen, Aachen, Germany