Yasushi Gohou’s scientific contributions

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (5)


Neuromorphic computing based on Highly reliable Analog ReRAM by filament control
  • Conference Paper

September 2019

·

11 Reads

·

1 Citation

T. Mikawa

·

R. Yasuhara

·

K. Katayama

·

[...]

·

T. Kakiage




Citations (3)


... In the field of emerging memories, the Resistive Random-Access-Memory (ReRAM) [6] constitutes an alternative for the development of high-density and ultra-low-power applications [7][8]. One of the most promising properties of the ReRAM devices is the capability of exhibiting multiple resistance states [9], enabling them to operate in a Multi-Level-Cell (MLC) regime [10], which is a property required for the development of in-memory computing systems [11][12][13]. Transition Metal Oxide (TMO) based resistive memories (OxRAMs) are considered one of the most viable solutions to implement artificial synapses for neuromorphic computing systems [14]. Typically, during a pulse set operation, the conductance of the device varies based on the pulse height, width, and number of pulses. ...

Reference:

Characterization of Switching Properties in ReRAM Devices by the Capacitance of the MIM Structure
Neuromorphic computing based on Analog ReRAM as low power solution for edge application
  • Citing Conference Paper
  • May 2019

... Many researchers have focused on exploiting the capabilities of emerging NVM (non-volatile memory) devices for analog computation in order to realize latency and energy benefits of 1000× 1 with respect to conventional graphics processing unit (GPU) and central processing unit (CPU) based computing for implementing largescale networks. [2][3][4] The key limitations of such methods are (i) the need for a reliable multi-state capability of the device in the optimal sensing range, 5 (ii) the requirement of selector devices to reduce the impact of sneak paths, and (iii) the need for linearity in multi-level cell (MLC) states along the sensing range. However, selector devices lead to increased on-chip area utilization and control signal routing complexities. ...

Reliability Issues in Analog ReRAM Based Neural-Network Processor
  • Citing Conference Paper
  • March 2019

... Tsing Hua University reported a fully integrated analog ReRAM-based chip with fully parallel multiplication accumulation (MAC) computing capabilities [ 192 ]. Panasonic Corporation reported a neural network processor using analog ReRAM [ 193 ]. International Business Machines (IBM) Corporation demonstrated a HERMES core based on PCRAM [ 194 ]. ...

A 4M Synapses integrated Analog ReRAM based 66.5 TOPS/W Neural-Network Processor with Cell Current Controlled Writing and Flexible Network Architecture
  • Citing Conference Paper
  • June 2018