Few-Shot Image Recognition with Device-Cloud Collaboration

Deep learning has shown great success in a variety of tasks with large amounts of labeled data in image classification. These achievements have relied on the fact that optimization of these deep, high-capacity models requires many iterative updates across many labeled examples. This type of optimization breaks down in the small data regime where we want to learn from very few labeled examples. In this project, we are interested in the few-shot learning problem. In particularly, we focus on two challenging scenarios which are 1) to deploy on both the device side (e.g., mobile phones) and the cloud side collaboratively, and 2) to deploy on end devices only. The intern will work closely with Huawei Canada to develop advanced algorithms and delivery prototypes.

Faculty Supervisor:

Jiangchuan Liu

Student:

Partner:

Huawei Technologies Canada Co Ltd (Burnaby, BC);Huawei Technologies Canada Co Ltd (Kanata, ON)

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

Simon Fraser University

Program: