Confined PCM-based Analog Synaptic Devices offering Low Resistance-drift and 1000 Programmable States for Deep Learning

W. Kim(IBM Research - Thomas J. Watson Research Center), M. BrightSky(IBM Research - Thomas J. Watson Research Center), Stefano Ambrogio(IBM (United States)), Fabio Carta(IBM Research - Thomas J. Watson Research Center), Nanbo Gong(IBM Research - Thomas J. Watson Research Center), M. Longstreet(IBM Research - Thomas J. Watson Research Center), Praneet Adusumilli(IBM Research - Thomas J. Watson Research Center), Takeshi Masuda(Ulvac (Japan)), Jin‐Ping Han(IBM Research - Thomas J. Watson Research Center), Gloria Fraczak(IBM Research - Thomas J. Watson Research Center), Hsinyu Tsai(IBM Research - Almaden), J. Bruley(IBM Research - Thomas J. Watson Research Center), Koukou Suu(Ulvac (Japan)), Robert L. Bruce(IBM Research - Thomas J. Watson Research Center)
Unknown
June 1, 2019
10.23919/vlsit.2019.8776551
Cited by 70

Related Papers

Equivalent-accuracy accelerated neural-network training using analogue memory
|Nature|2018|1.1k
Recommended Methods to Study Resistive Switching Devices
|Advanced Electronic Materials|2018|649
Emerging neuromorphic devices
|Nanotechnology|2019|312
Neuromorphic Learning and Recognition With One-Transistor-One-Resistor Synapses and Bistable Metal Oxide RRAM
|IEEE Transactions on Electron Devices|2016|250
Unsupervised Learning by Spike Timing Dependent Plasticity in Phase Change Memory (PCM) Synapses
|Frontiers in Neuroscience|2016|246