Machine Learning for signal processing and pulse sequence design - ON-520

Project type: Research
Desired discipline(s): Engineering - computer / electrical, Engineering, Computer science, Mathematical Sciences, Mathematics
Company: Foqus Technologies Inc.
Project Length: 6 months to 1 year
Preferred start date: 10/01/2021
Language requirement: English
Location(s): Toronto, ON, Canada; Canada
No. of positions: 1-2
Desired education level: PhDPostdoctoral fellow
Search across Mitacs’ international networks - check this box if you’d also like to receive profiles of researchers based outside of Canada: 
No

About the company: 

Foqus provides solutions based on Quantum Information algorithms and Machine Learning techniques to enhance the sensitivity of Magnetic Resonance Technologies such as MRI and NMR.

Describe the project.: 

There are two objectives. The first one is for the enhancement of the pulse sequence that implements our quantum algorithms. Currently, we are using the optimal control theory. The first part of the project is to use machine learning (probably reinforcement learning) to enhance the pulse sequence.

The second part of the project is to enhance the response signal after the application of the pulse sequence. The task is to develop machine learning models for enhancing the signal. This involves amplification of the signal, suppressing the noise, and reconstruction of a partial signal.   

Required expertise/skills: 

The ideal candidate is a PhD student or postdoctoral fellow in Computer Science.

Expertise:

  • Deep learning (convolutional NN, Recurrent NN).
  • Reinforcement learning.
  • (Optional) Experience with signal processing using ML techniques.