preview all symposia

INFORMATION AND COMMUNICATION TECHNOLOGIES

F

Novel Approaches for Neuromorphic Computing: Materials, Concepts and Devices

Energy consumption and data storage limit the development of modern electronics. This motivates the search for alternative ways of information processing. Bioinspired electronics based-on non-volatile memory devices is most promising. Various materials, technologies and computational schemes can be used for implementing artificial synapses and neurons in so-called neuromorphic networks.

Scope

Neuromorphic engineering exploiting non-volatile memory (NVM) devices, or in general memristive systems), has enormous potential for highly energy-efficient cognitive electronics being clearly superior to state-of-the-art computing architectures in terms of data- and energy consumption. To mimic the role of synapses in the nervous systems and the stochastic and non-linear characteristics of neuronal units, a large number of materials and device structures are investigated for neuromorphic systems. Furthermore, a variety of different computational approaches is already presented taking into account the special properties of those devices and various architectures, also exploiting hybrid CMOS-NVM systems. For example, the binary On/Off switching behavior, as explored in Conductive-Bridging RAM (CBRAM) or Magnetic RAM devices can be used to emulate synaptic functionality. Phase Change Memory (PCM) or Resistive-RAM (RRAM), are as well highly employed to reconstruct phenomenological substitutes of synapses for neural networks. Furthermore, the inherent stochasticity and multilevel, or analog, capability of some NVM devices are used to emulate synaptic learning and memory processes. However, in order to realize powerful neuromorphic electronics, an interdisciplinary approach is needed that ranges from the investigation of material properties, through the realization of functional devices, to novel computing and circuit schemes. In order to maintain this interdisciplinary, an intensive exchange of scientists with very different expertise is required. This includes experts in the field of novel materials, analytics and measurement methods, as well as device experts, developers of computational and biological modeling, circuit designers and experts in the field of complex dynamic systems. Therefore, the symposium seeks to provide the framework for an interdisciplinary exchange of scientists from those various fields.

Hot topics to be covered by the symposium

  • Physics and technology of memristive nanomaterials and devices
  • Stochastic phenomena in memristive materials
  • Engineering of different technologies devices (RRAM, CBRAM, MRAM, FeRAM, PCM) for spiking neural networks
  • Devices to emulate the synaptic or neuron functionality
  • Emerging materials and concepts for neuromorphic computing: 2D materials, organic synaptic devices, optical switching
  • Design and modelling of neuromorphic computing systems
  • Bio-hybrid network
  • Organic Electronics for Neuromorphic Computing
  • Photonic neural networks

No abstract for this day

No abstract for this day

No abstract for this day

No abstract for this day


Symposium organizers
Beatriz NOHEDAUniversity of Groningen

Zernike Institute for Advanced Materials, Nijenborgh 4, 9747AG Groningen, The Netherlands

b.noheda@rug.nl
Christian WENGERIHP GmbH

Leibniz Institut fuer innovative Mikroelektronik, Im Technologiepark 25, 15236 Frankfurt/Oder, Germany

wenger@ihp-microelectronics.com
Martin ZIEGLERTU Ilmenau

Microelectronic and Nanoelectronic Systems, Gustav-Kirchhoff-Str., 98684 Ilmenau, Germany

martin.ziegler@tu-ilmenau.de
Sabina SPIGACNR-IMM

Unit of Agrate Brianza, via C. Olivetti 2, 20871 Agrate Brianza (MB), Italy

sabina.spiga@mdm.imm.cnr.it