Brainmorphic Computing Hardware
update：2021-07-12
Features
We will develop a brainmorphic computing hardware, which realizes the brain-specific functions such as conscious/sub-conscious process, self, selective attention, and so on, by directly using inherent physics and dynamics of constituent devices. The resulting hardware would be small, efficient, high-performance. Some examples include the chaotic neural network reservoir, optimization through high-dimensional complex dynamics, and neural network composed of spin-orbit torque nano-devices.The resulting hardware is suitable for the edge AI which learns users’ personal behavior. Examples include watching service devices embedded in hearing aids or dental implants, which monitor and learn personal cardiac and brain-wave signals or saliva ingredients, to detect abnormal situations.
Targeted Application(s)/Industry
Edge AI devices, especially for peri-personal space), Time-series processing (prediction, recognition, and categorization), Online real-time learning.Researchers
Research Institute of Electrical Communication
HORIO Yoshihiko
, Professor
Ph.D
Keywords
Related Information
K. Fukuda, Y. Horio, T. Orima, K. Kiyoyama, and M. Koyanagi, "Cyclic reservoir neural network circuit for 3D IC implementation," Nonlinear Theory and Its Applications, IEICE, vol. 12, no. 3, pp. 309-322, DOI: 10.1587/nolta.12.309, July 1, 2021. (Invited Paper)
A. Kurenkov, S. DuttaGupta, C. Zhang, S. Fukami, Y. Horio, and H. Ohno, "Artificial neuron and synapse realized in an antiferromagnet/ferromagnet heterostructure using dynamics of spin-orbit torque switching," Advanced Materials, 1900636, DOI: 10.1002/adma.201900636, April, 2019.
Y. Horio, "A brainmorphic computing hardware paradigm through complex nonlinear dynamics," in Understanding Complex Systems, V. In, P. Longhini, and A. Palacios, eds., Springer, ISBN 978-3-030-10891-5, Chapter 5, pp. 36-43, DOI: 10.1007/987-3-030-10892-2_5, 2019.
Prev
List
Next