Efficient edge computing benchmarking for AI-driven applications- ON-306

Desired discipline(s): Engineering - computer / electrical, Engineering, Engineering - mechanical, Engineering - other, Computer science, Mathematical Sciences
Company: LEI Technology Canada, a division of Lanner Electronics
Project Length: 4 to 6 months
Preferred start date: As soon as possible.
Language requirement: English
Location(s): Mississauga, ON, Canada
No. of positions: 1
Preferred institutions: Algoma University, Brock University, Carleton University, Concordia University, École Polytechnique de Montréal, HEC Montréal, Lakehead University, Laurentian University, McGill University, McMaster University, Nipissing University, Ontario Tech University, Polytechnique Montréal, Queen's University, Royal Military College of Canada, Ryerson University, Trent University, Université de Montréal, Université de Sherbrooke, Université du Québec : Institut national de la recherche scientifique, Université du Québec à Chicoutimi, Université du Québec à Montréal, Université du Québec à Rimouski, Université du Québec à Trois-Rivières, Université du Québec en Abitibi-Témiscamingue, Université du Québec en Outaouais, Université Laval, University of Guelph, University of Montreal, University of Ontario Institute of Technology, University of Ottawa, University of Toronto, University of Waterloo, University of Western Ontario, University of Windsor, Wilfrid Laurier University, York University

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: 

Lanner Electronics Inc is a world leading provider of design, engineering and manufacturing services for advanced network appliances and rugged applied computing platforms for system integrator, service providers and application developers.

Over the past ten years, Lanner has dedicated itself in supplying millions of state-of-the-art industrial computing platforms that have been deployed worldwide as Intelligent Edge Appliances to perform analysis, connectivity and storage of the data generated within proximity of sensor and devices. Our comprehensive rugged computing platforms have met various specifications and standards required by the mainstream Intelligent Edge verticals. Today, we have comprehensive product lines designed and manufactured for industrial AI, machine vision, transportation, intelligent video analytics, substation automation and OT network security.

Please describe the project.: 

AI-based Edge Computing technology is finding its way into various application scenarios. As outlined by the International Electrotechnical Commission (IEC), Edge intelligence pushes processing for data intensive applications away from the core of the cloud to the edge of the network. Reallocating computing and analytics to Intelligent Edges can reduce the latency, cost and security arise from communications between the data source and the data centers.

Lanner develops a range of Intelligent Edge products that address industrial computing needs in the areas of industrial AI, automation, transportation, and substation and OT security. The proposed project will explore the following two general areas:

1. Edge Computing alignment with AI hardware accelerators and AIOps technologies (mainly NVIDIA and Intel based), involving:

  • Compatibility test of AI accelerator solution with multiple Lanner commercial off the shelf (COTS) products. For example, NVIDIA T4 PCIe*16 card and Intel Movidius Myriad X mPCIe card with LEC-2290.
  • Functional test of inference stack: AI optimized libraries, graph compilers and formats, and multiple ML frameworks, models and datasets with Lanner Edge AI appliances.

2. Lanner Edge AI appliance performance (deep learning inference and training) benchmarking on public platform, involving:

  • Submission of selective IEC Edge AI appliances to MLPerf Inference benchmarking suite for measuring how fast systems can process inputs and produce results using a trained model.

Required expertise/skills: 

  • Mobile, pervasive, cloud, and edge computing
  • Edge computing algorithms and systems
  • Distributed machine learning