抖阴APP导航

It wasn鈥檛 that long ago when surfing the internet involved bleeps and bloops while users hoped others in the house didn鈥檛 pick up the phone.

The world of connectivity and technology has changed dramatically over the last two decades, with machine learning and artificial intelligence becoming more than just a science-fiction story.

, PhD, has dedicated his research to understanding and improving these technologies to make them more accessible, inexpensive and, believe it or not, smarter.

His work on developing more-efficient deep neural networks (DNNs) 鈥� the technology that allows us to search the photos on our mobile devices, ask questions of smart assistants or use machine translation to understand a foreign language 鈥� has captured the attention of many in the technological world and has been widely used in enabling artificial intelligence (AI) applications like facial recognition on smartphones.

Ioannou was one of only 20 researchers in Canada to be awarded a DND/NSERC Discovery Grant Supplement, a joint initiative between the Department of National Defence (DND) and the Natural Sciences and Engineering Research Council of Canada (NSERC), and, most recently, he was among .

Exponential growth

Winning the awards came as a huge surprise to Ioannou, who joined the University of Calgary鈥檚 Schulich School of Engineering as an assistant professor in the in September 2021.

鈥淭hese awards were the first two research proposals I wrote,鈥� he says. 鈥淲riting proposals for grants and awards such as these isn鈥檛 something I ever had the experience of doing before, and I wasn鈥檛 sure I鈥檇 be successful, especially in a highly competitive field of candidates.鈥�

With regards to the Amazon award, Ioannou says it鈥檚 an honour to be included among some of the world鈥檚 leading researchers.

He says it wasn鈥檛 that long ago that he was a high school student taking part in a computer science summer outreach program from the University of Toronto.

After hearing a talk from deep learning pioneer , PhD, Ioannou was hooked on computer vision and neural networks, leading to a post-secondary journey that concluded with completing his PhD at the University of Cambridge in 2018.

鈥淎s someone who was the first in my family to attend university, never mind graduate school, research wasn鈥檛 even on my radar before that talk,鈥� he says. 鈥淭his really goes to show the impact that university outreach programs can have.鈥�

An exercise in networking

The Amazon award was given to Ioannou for his work on machine learning algorithms and theory, regarding what is called 鈥渃atastrophic forgetting.鈥�

He says DNNs can鈥檛 learn like humans do, like learning different tasks easily, and don鈥檛 easily learn or adapt to new data while interacting with the real world.

鈥淚t means we typically have to use multiple models for different tasks, even when those tasks are very related,鈥� Ioannou says. 鈥淲e can鈥檛 put a neural network in some application and expect it to be able to adapt to new data or conditions that it encounters.鈥�

These fundamental problems, along with the increasing costs associated with improving the technology, has led to the development of sparse neural networks (SNNs).

鈥淭hese networks present potential solutions to several of these problems, and addressing any one of these problems could have a large real-world impact on the increasingly large number of applications of DNNs,鈥� Ioannou says.

Focused on the future

As the technology continues to evolve, Ioannou has no shortage of focuses for his research in the future.

He says it鈥檚 very expensive to learn and apply a DNN to new tasks, so he would like to find ways to make it more accessible.

Ioannou says the support he has received allows him to think bigger about his projects and accelerate his research program.

鈥淧erhaps most importantly, it allows me to increase the number of graduate and undergraduate students I can train in what is a very in-demand field of expertise that will drive Alberta and Canada鈥檚 future,鈥� he says.