Neuromorphic Engineering

from Wikipedia, the free encyclopedia

Neuromorphic Engineering (engl.) Is a new interdisciplinary direction of biologically inspired physics, mathematics, computer science and engineering to develop neural systems and improve artificial intelligence . This is used to develop applications such as machine vision, " head-eye systems ", machine hearing comprehension, and autonomous robots.

Research priorities

  • The European-wide interdisciplinary research project FACETS aims to develop novel biology-inspired computing concepts that differ fundamentally from conventional IT systems based on the Turing model. The focus of the project is the development of a neuromorphic, analog hardware architecture in VLSI technology for the simulation of macroscopic parts of the cerebral cortex . The project is funded by the European Commission under the 6th Research Framework Program .
  • Kwabena Boahen , a neuro-engineer at Stanford University , wants to use neuromorphing to create a circuit with around one million artificial neurons made of silicon.
  • The Korea Advanced Institute of Technology has been working with colleagues since 1998 on a lifelike replica of human sensory organs including signal processing called the Artificial Brain . The project is led by Soo-Young Lee . Two cameras are movably suspended here. The signals from the cameras are processed by circuits that are based on the human retina. Two microphones are connected to improved cochlear chips . As a result, part of the information processing is outsourced to the sensors, which relieves the downstream information processing entity. In addition, the information can be filtered according to attention . The decision which information is relevant here is intended in further steps by a Emotions - function can be determined.
  • The Blue Brain project at the Swiss Federal Institute of Technology in Lausanne simulates the interaction of around 10 6 completely realistic neurons of a rat brain on a high-performance computer. The project is led by Henry Markram .
  • IBM Cognitive Computing maintains the SyNAPSE program , in which a cat's brain is simulated with about 10 9 not-completely-realistic neurons. The project is led by Dharmendra Mohda .
  • The Brain Corporation , headed by Eugene M. Izhikevich, simulates about 10 10 neurons, which is the size of the human brain . However, these neurons are randomly linked and have no specific function.
  • The NIH maintains the Human Connectome Project , which studies the large-scale nerve connections in the brain.
  • The University of Waterloo is working on the "functional brain model" Spaun , which is based on the Nengo framework and consists of 2.5 · 10 6 neurons and can perform several different tasks without changes.

See also