Personal tools

Difference between revisions of "Optical Weights for Photonic Neural Networks"

From iis-projects

Jump to: navigation, search
 
Line 56: Line 56:
  
 
[[#top|↑ top]]
 
[[#top|↑ top]]
 +
 +
<!--
 +
 +
COPY PASTE FROM THE LIST BELOW TO ADD TO CATEGORIES
 +
 +
GROUP
 +
[[Category:Digital]]
 +
[[Category:Analog]]
 +
[[Category:Nano-TCAD]]
 +
[[Category:Nano Electronics]]
 +
 +
STATUS
 +
[[Category:Available]]
 +
[[Category:In progress]]
 +
[[Category:Completed]]
 +
[[Category:Hot]]
 +
 +
TYPE OF WORK
 +
[[Category:Semester Thesis]]
 +
[[Category:Master Thesis]]
 +
[[Category:PhD Thesis]]
 +
[[Category:Research]]
 +
 +
NAMES OF EU/CTI/NT PROJECTS
 +
[[Category:UltrasoundToGo]]
 +
[[Category:IcySoC]]
 +
[[Category:PSocrates]]
 +
[[Category:UlpSoC]]
 +
--->

Latest revision as of 14:42, 4 September 2019

Short Description

At IBM Research – Zurich the Neuromorphic Devices & Systems group is working on an implementation of a neural network based on integrated silicon photonic technology. Recently we developed integrated, multi-level optical synaptic elements. These elements will be embedded in an optical neural network. In a first part of the Master thesis, you will address the characterization of such weighting elements. During this task, you will measure the device transmission, set up and perform the electro-optical measurements and analyze your measurement results. In a second task, you will design an interface between hardware and software to control multiple of these elements in parallel. Therefore, designing of electrical connections, programming of microcontrollers, as well as setting up an interfacing software will be required.

The Big Picture-Neuromorphic Computing

Artificial intelligence (AI) is the ability to perform tasks that are generally associated with intelligent beings. Recently, bio- and neuro-inspired (neuromorphic) algorithms have attracted considerable attention with their ability to extract structure and knowledge from huge unstructured data sets by relying solely on limited domain expert knowledge. To execute these new algorithms efficiently at the large scale required in datacenters or, for example, to interpret sensor data locally in embedded, very low-power devices, we need novel neuromorphic compute architectures and hardware. To compute and train neuromorphic and AI algorithms, digital processors (CPUs, GPUs, TPUs, FPGAs and ASICs) are used today. One promising alternative to the large, costly and power-hungry digital logic is analog computing, where computationally expensive operations are offloaded to specialized accelerators comprising analog elements with the promise to accelerate existing schemes by factors of 1000 to 10,000. Non-volatile synaptic elements are at the core of such novel computing systems.

The Environment-IBM Research Zurich

IBM Research GmbH, Zurich Research Laboratory (ZRL), with approximately 350 employees, is a wholly-owned subsidiary of the IBM Research division with headquarters at the T.J. Watson Research Centre in Yorktown Heights, NY, USA. ZRL, which was established in 1956, represents the European branch of IBM Research. At ZRL scientific and industrial research is conducted in five scientific and technical departments, in particular in the: Science and Technology department.

Throughout the history of this department, scientists have made major contributions to the advancement of knowledge in solid-state physics, stimulated by problems relevant to technology. Today research focuses on different areas of technological significance, such as photonics and optoelectronics, CMOS and post-CMOS, micro fabrication, packaging and life sciences. This effort is supported by the newly opened Binnig and Rohrer Nanotechnology Centre, offering state of the art micro fabrication yet with the required flexibility for any research at the frontier between industrial and academic research.

With more than 100 students and young researchers (including Master and PhD students, as well as postdoctoral researchers), our laboratory offers a dynamic and international environment for excellent science and provides a unique opportunity to extend your research skills. The laboratory equipment for the project defined in this Master thesis is state-of-the art, and allows to perform novel and exciting experiments.


Status: Available

Looking for 1 Master student
Interested candidates please contact: Dr. Stephan Abel
ETH Contact: Mathieu Luisier

Prerequisites

We are seeking a candidate with a strong interest in integrated optics as well as basic knowledge of microcontroller programming, object-oriented programming and circuit design. You should be enrolled as a student at ETH Zurich. For this master project you should be available for a period of at least 6 months starting in Fall 2018.


Professor

Mathieu Luisier

↑ top