Kategorie: ‘Mikro- und Nanoeletronik’
RWTH and regional partners successful in the “Clusters4Future” ideas competition
Future clusters “NeuroSys” and “Hydrogen” to receive up to 90 million euros in funding from the BMBF
The German Federal Ministry of Education and Research (BMBF) has now published the winners of the “Clusters4Future” ideas competition. The BMBF is funding the research with up to 90 million euros. “Clusters4Future” is part of the German government’s High-Tech Strategy 2025. The open-topic competition focuses on regional innovation networks that combine the strengths of the players, tap into emerging fields of innovation and develop solutions for the challenges of the future.
“NeuroSys – Neuromorphic Hardware for Autonomous Artificial Intelligence Systems”
In addition to RWTH, the Forschungszentrum Jülich, AMO GmbH, IHK Aachen, the companies AixACCT Systems GmbH, AIXTRON SE, AppTek GmbH, ELMOS Semiconductor SE, RWTH Innovation GmbH and STAR Healthcare Management are involved in NeuroSys. In addition, the start-ups AiXscale Photonics UG, Black Semiconductor GmbH, Clinomic GmbH and Gremse-IT GmbH are involved. Professor Max Lemme from the Chair of Electronic Components and Managing Director of AMO GmbH will coordinate the work. The goal is the development of neuromorphic hardware for artificial intelligence applications and thus a technological independence for Germany and Europe. The Federal Ministry of Education and Research is providing up to 45 million euros for this purpose.
In Europe, there are only a few global corporations in the hardware and software sector. Technological independence is of strategic importance, as artificial intelligence will be the building block for the next global stage of development. However, not only future economic growth depends on this key technology, but also the management of major societal challenges such as climate change, health, work or mobility. At the same time, artificial intelligence (AI) brings new challenges. For example, training large neural networks based on modern graphics processing units (GPUs) with deep-learning methods causes high CO2 emissions, which further exacerbate the climate problem. GPU-based neural networks are therefore ecologically unsustainable.
Resource-saving neuromorphic hardware that makes neural networks more efficient and includes data security as a design component is therefore becoming the key to the widespread use of AI. This is especially true for areas of application in autonomous vehicles, medical technology and sensor networks for intelligent production or urban regions. Neuromorphic systems are modelled on the two basic building blocks of the human brain, the neurons and the synapses. By integrating new materials with specific properties, they can ideally perform resource-saving on-site processing of data by integrating new materials with certain properties. This is summarised under the keyword “memristive” – from the English “memory” for storage and “resistor” for electrical resistance.
Scientists at RWTH and Forschungszentrum Jülich have already been able to demonstrate the functionality of neuromorphic devices made of memristive materials. However, there are no pilot lines or production capacities worldwide for manufacturing or integrating neuromorphic chips on an industrial scale. Also, the system of hardware, design, algorithms and application-driven software must work together to exploit the major advantages of neuromorphic hardware. What is needed, therefore, is a paradigm shift with the opportunity to take a leading position in this new technology. NeuroSys wants to develop the decisive prerequisites here.
In addition to economic success, aspects such as the social benefits and ethics of artificial intelligence must be taken into account. These socio-economic framework conditions are essential for new technologies, especially with such a potential reach. They are therefore being researched in NeuroSys, also in order to develop recommendations for action for society and politics.
“The Future Cluster is a great opportunity for the Aachen-Jülich region, especially in connection with the structural change in the Rhenish Revier. We are stepping up to transfer excellent science into companies and start-ups in the region. Our vision is to set up a production line in the Aachen region. There, the co-integration of neuromorphic functions through new materials into conventional silicon technology will then take place” – Professor Lemme.
We also congratulate the Institute for Combustion Engines, since in addition to “NeuroSys”, the seven funded clusters include the future cluster “Hydrogen”, which will also be coordinated by RWTH Aachen University in the coming years. RWTH Aachen University and Forschungszentrum Jülich were applicants for the future cluster “Hydrogen”. So far, 24 institutes of the two research institutions are involved, in addition to 47 industrial partners and 16 other organizations.
Neuromorphic Computing
DFG grant for “Memristive Devices Toward Smart Technical Systems”
The German Research Foundation (DFG) is funding five projects under the priority program “Memristive Devices Toward Smart Technical Systems” with the participation of members of the Faculty of Electrical Engineering and Information Technology at RWTH Aachen University. Four of them are projects at the chair of Prof. Rainer Waser IWE2 and the Peter Grünberg Institute of the Forschungszentrum Jülich. A further project was approved in the teaching and research area of Prof. Regina Dittmann “Technology of Oxide Electronics” also at the Peter Grünberg Institute.
The funding for the five Jülich-Aachen projects amounts to approx. 1.2 million EURO for the duration of the priority program of 3 years. Within the framework of the various projects, Faculty 6 will develop memristive components for use in novel energy-efficient computer structures or for intelligent sensor applications for the future Internet of Things in cooperation with other research institutions such as TU Dresden, TU Chemnitz, the Karlsruher Institut für Technologie (KIT), the Helmholtz-Zentrums Berlin, TU Berlin and the NMI – Natural and Medical Sciences Institute – at the University of Tübingen and the Groningen Cognitive Systems and Materials Center (CogniGron
About the projects:
In the project “Memristive Time difference encoder (MemTDE)” the group of Mrs. Dittmann and the Groningen Cognitive Systems and Materials Center (CogniGron) are working on the development of a memristor-based intelligent electronics for processing sensor signals for the Internet of Things. This is intended to process the collected information on site instead of transmitting it wirelessly using a lot of energy.
In the “Hybrid MEMristor-CMOS Micro Electrode Array bio-sensing platform (MEMMEA)” project, the partners of PGI-7, the Helmholtz Center Berlin, the TU Berlin and the NMI – Natural and Medical Sciences Institute – at the University of Tübingen are striving to develop sensors that can directly record the activity of biological neurons. These sensors based on memristor-CMOS hybrid circuits enable direct on-chip signal processing and open up a new field of biological signal processing.
In the project “Domino Processing Unit: Towards Novel High Efficient In-Memory-Computing (MemDPU)” the partners of PGI-7 and the Chemnitz University of Technology are working on a novel computing unit, the Domino Processing Unit (DPU). In contrast to the conventional von Neumann architecture computing unit, this DPU enables computing directly in memory. With the DPU, the high energy consumption is saved through communication between the memory and the computing unit.
In the project “Universal Memcomputing in Hardware Realizations of Memristor Cellular Nonlinear Networks (Mem2CNN)” the partners of PGI-10, PGI-7 and TU Dresden are pursuing the development of memristive cellular neural networks. These networks enable the direct processing of video signals, for example in the form of edge detection for pattern recognition. Thus, visual data could be processed in real-time.
In the project “Robust Compute-in Memory using Memristors : ROBCOMM”, the partners of IWE 2, PGI-7 and Karlsruhe Institute of Technology (KIT) are working on the development of reliable, efficient circuits based on memristive components that enable a Computation-in-Memory (CIM) architecture. The CIM architecture allows to efficiently perform complex computational operations such as vector-matrix operations or to directly solve large systems of equations.
