Deep learning at the service of the fashion industry - QC-113

Preferred Disciplines: Physics, Mathematics, Computer Science, Software engineer or similar. PhD or Postdoc level.
Company: N/A
Project Length: 4 to 6 months
Desired start date: As soon as possible
Location: Montreal, Quebec
No. of Positions: 1
Preferences: Any University with labs doing Machine Learning/Deep Learning. No Language preferences.

About Company:

The company is one of the city's fastest-growing tech companies. We are currently seeking Research Scientists to join our team as we are launching new projects in applied artificial intelligence and research in artificial intelligence. Our current team consists of scientists who are committed to balancing intensive scientific research with industry product expertise. We are cultivating an environment that gives our staff the resources to conduct research that can be presented through papers in academia and contributions to scientific journals. In addition, it is our aim to attend key conferences in our network to contribute to new ideas, discoveries and more.

At the moment, we have research partners in multiple different industries: health, sports, social media, communications, and security. From these partners, we are accruing key data points to identify real-world problems to be solved with AI-driven products and technologies.

We are passionate about the products we create and the partners we work with. We’re committed to staying ahead of the curve by constantly learning what’s new in the industry. This makes for a fun, dynamic and exciting work environment that’s in constant evolution. Read on to see if your experience fits the bill.

Project Description:

Expand  the capabilities of  our proprietary Image Retrieval engine based on Deep Learning techniques. This image retrieval engine, have the property of understanding human taste and provides users with valuable recommendations. This engine is currently deployed in several e-commerce sites. 

Background and required skills

 

Research Objectives/Sub-Objectives:

Objective 1: Use GAN to translate designer sketches into fashion pieces based on current trends

Objective 2:  Given a set of fashion designs being able to determine which one match better with the current trends

Objective 3: Research and implement a method to extract relevant pixels/regions on the picture and restrict features extraction to that region

Note: in all cases, the intern will work in the Data Science part of the project, leaving final implementation and software engineering tasks to our staff. 

Methodology:

  • GAN, Convolutional Neural Networks, Fine tuning/transfer learning, auto-encoders, visualizations techniques, semantic segmentation, attention mechanism.

Expertise and Skills Needed:

  • Experience in Deep Learning, Machine Learning, NLP/NLU and/or computer vision
  • Knowledge of Python
  • Comfort using (or willing to do a quick ramp up) Tensorflow
  • Enthusiasm for working on a team and a passion for analyzing and finding solutions

 

For more info or to apply to this applied research position, please

 

  1. Check your eligibility and find more information about open projects.
  2. Complete this webform. You will be asked to upload your CV. Remember to indicate the title of the project(s) you are interested in and obtain your professor’s approval to proceed!
  3. Interested students need to get the approval from their supervisor and send their CV along with a link to their supervisor’s university webpage by applying through the webform or directly to Jean-Philippe Valois  at, jpvalois@mitacs.ca

Program: