#Canada#Toronto#Tech#Artificial Intelligence#Vector Institute

Vector Institute: researching the future

Knackles
|Feb 5|magazine8 min read

As Canada experiences a rush of investment within the AI sector, the Vector Institute for Artificial Intelligence is leading the pack.

Foreseeing the possibility to improve the lives of ordinary Canadians, the Toronto-based institute was founded in 2017 with the vision of driving excellence, leadership, knowledge, creation and usage of artificial intelligence (AI).

But what are the basics of AI and machine learning?

Generally, machines are provided known ‘inputs’ and their associated ‘outputs’ and then give a predicted outcome for a set of new ‘inputs’. The machine’s ultimate goal is, based on previous results, to keep making predictions until the error margin (called the ‘cost function’) between the predicted and actual output is as low as possible. 

The machine learning process is based around this premise.

Although the technology being developed by the Institute is far from ordinary, the day-to-day operations and concerns of the organisation are the same as any business.

SEE ALSO:

“We’re here to help superstar researchers but our main goal is to strengthen the region,” said Garth Gibson, CEO, in an article with Science Business. “A strong workforce, a compelling place to work, a good tax law and a sense of competitiveness among your people. We have all of that.”

In the business of breakthroughs

News and updates on the Institute’s website are uploaded regularly, with the latest breakthroughs or projects showing potential outlined in scientific detail. Recent projects include:

New neural networks: The Institute has developed new models with the enhanced potential to predict patterns in medicine, finance and even human genetics. Using backpropagation, computers are able to solve a problem using ‘gradient descent’. 

Uncovering injuries: Utilising machine-learning, a paper written by researchers Boshra, Dhindsa and Boursalie et al, offers the opinion that computers might be able to recognise symptoms of concussion in patients years after their original injury. 

The research team’s algorithm can allegedly achieve an 81% accuracy rate - a percentage which eclipses current clinical capabilities. In addition, the algorithm could determine which specific part of the brain was affected.

For more information on business topics in Canada, please take a look at the latest edition of Business Chief Canada.

Follow Business Chief on LinkedIn and Twitter