Job Detail

Master's student position Image-based materials defect classification using machine learning algorithms

Inseriert am: 14.01.2019

Ref. 2019-02


Project description



Defective (left) and non-defective (right)
III–V crystalline layer
In the world of nanoscale electronics and photonics, material defects can be detrimental to device performance. However, some defects can also enhance certain properties, whence the need to accurately characterize materials grown in our laboratory.


In the Material Integration and Nanoscale Devices group at IBM Research – Zurich, we are more specifically focusing on the growth of new materials analyzed by transmission electron microscopy (TEM). Obtaining high-resolution images allows any crystalline defect to be detected and identified at the nanoscale level. However, this requires substantial manual work and experience. In that respect, we believe that machine learning could be a key enabler to classify defects accurately and, in turn, allow a better understanding of how they form and impact the fabricated devices. Furthermore, such AI-based assistance systems will permit us to classify a much larger number of TEM images than what is possible with the current manual classification by experts.


Our multidisciplinary teams at IBM Research – Zurich are developing new materials as well as computational techniques that extract, quantify and classify visual patterns that differentiate between high-performance and defective devices.


In collaboration with the imaging team at IBM Research – Zurich, the goal of the proposed master project is to combine image processing, pattern recognition, deep learning and material domain expertise to develop a classifier that ingests high-resolution TEM images, learns to recognize defective from non-defective crystalline areas and, at a latter stage, is also able to distinguish among the defects.


The Master student project will focus on:



  1. Generating TEM images of defects in III–V materials

  2. Developing and testing a classification framework of computational defects

  3. Determining appropriate material growth conditions for high-performance nanoscale devices.


The minimum duration for this Master’s thesis research is 6 months.


Please note that this Master’s thesis research project at IBM provides an opportunity to carry out your thesis in a world-class research environment, but it is not funded. The student should preferably already be enrolled in a Swiss educational institute such as ETHZ.


Requirements


We are looking for a highly motivated, creative and independent student ofComputer Science or Electrical Engineering who has a good background in image processing, pattern recognition and deep learning. Additional knowledge of materials science will be an advantage.


Diversity


IBM is committed to diversity at the workplace. With us you will find an open, multicultural environment. Excellent flexible working arrangements enable both women and men to strike the desired balance between their professional development and their personal lives.


How to apply


Please send your application to Marilyne Sousa and Dr. Maria Gabrani.

Details