Personal tools

Ferroelectric Memristors for Artificial Neural Networks (IBM-Zurich)

From iis-projects

Revision as of 10:47, 23 June 2021 by Emborasa (talk | contribs) (Prerequisites)
Jump to: navigation, search

Short Description

At the Neuromorphic Devices and Systems laboratory in IBM Research, we have an opening for a Master research project:

Electrical characterization: You will set-up an experiment for the characterization of single devices and crossbar arrays of memristors, and process the data.
Research: you will learn about electrostatic field-effects and electron tunneling, to understand the conduction mechanisms in the characterized devices.
One of the outcomes of the project is to provide guidelines for the optimization of the technology

The Big Picture

As they scale up in size, training neural networks demand increasing time and energy. Therefor efforts have been put in developing dedicated analog hardware accelerators based on memristors. Key requirements include a suitable non-volatile resistance range, continuous linear resistance tuning and symmetric switching. Ferroelectric materials are promising candidates for memristive applications, as they have an intrinsic switchable and non-volatile property: the ferroelectric polarization.

Status: Available

Looking for 1 Master student. Prior to the Master thesis, you are welcome to carry a semester project in our group, during which you will learn about processing.
Interested candidates please contact: Laura Bégon-Lours

Prerequisites

Master student interested in electrical engineering, materials sciences, programming.


Mathieu Luisier

↑ top