Back to Search
Start Over
Mixed-mode Magnetic Tunnel Junction-based Deep Belief Network
- Source :
- 2019 IEEE 19th International Conference on Nanotechnology (IEEE-NANO).
- Publication Year :
- 2019
- Publisher :
- IEEE, 2019.
-
Abstract
- We present a mixed-mode magneto tunneling junction (m-MTJ)-based Deep Belief Network (DBN). DBNs are unsupervised learning models, suitable for recognition and clustering. m-MTJ is a three-terminal magnetic device with probabilistic free layer switching controlled by the simultaneous actions of voltage-controlled magnetic anisotropy and spin-transfer torque. While DBNs achieve high prediction accuracy even with highly imprecise single-bit weights, the key complexity lies in their activation functions which are stochastic. Using an m-MTJ, we present a novel low area/power DBN neuron with stochastic activation function. We discuss an in-memory computing architecture that allows forward and backward flow of learning dynamics and online learning. Our design achieves ~88.80% accuracy for digit recognition in MNIST even under the worst case variability in nanoscaled m-MTJs.
- Subjects :
- Computer Science::Machine Learning
010302 applied physics
Restricted Boltzmann machine
Computer science
Activation function
Probabilistic logic
02 engineering and technology
021001 nanoscience & nanotechnology
01 natural sciences
Power (physics)
Deep belief network
0103 physical sciences
Unsupervised learning
Hardware_ARITHMETICANDLOGICSTRUCTURES
0210 nano-technology
Cluster analysis
Algorithm
MNIST database
Subjects
Details
- Database :
- OpenAIRE
- Journal :
- 2019 IEEE 19th International Conference on Nanotechnology (IEEE-NANO)
- Accession number :
- edsair.doi...........46f9df03d0e7c2e57731640ec93db64d
- Full Text :
- https://doi.org/10.1109/nano46743.2019.8993914