New UTM professor’s computer science research definitely fashionable
Sanja Fidler is undoubtedly one of very few computer scientists whose research has been popularized in fashion magazines worldwide.
Fidler, an assistant professor of math and computer science at U of T Mississauga, gained renown last summer, along with St. George campus colleague, Raquel Urtasun, for developing an algorithm that assessed an outfit to determine whether or not it was considered fashionable in a particular locale and offered suggestions on how to improve it.
Fidler, a native of Slovenia, did a post-doctoral fellowship at U of T in 2011 before joining the faculty on the St. George campus in January 2014. Her appointment to U of T Mississauga took effect in July 2016, and she will teach her first course on campus in January 2017: Introduction to Image Understanding, a course that is new to the UTM campus. In future terms, she will offer Introduction to Machine Learning and a graduate seminar on topics relating to computer vision.
“I am looking forward to working with the students at UTM,” she says. “I am meeting with some students there this week to see if there is interest in doing a project together.”
With an undergraduate degree in mathematics and a PhD in computer science, Fidler specializes in computer vision and in the interplay between language and vision as it relates to machine learning.
“I employ machine learning all the time to make algorithms work,” Fidler says. “Machine learning ties all the modalities – computer vision, language and action – together.”
Much of the research Fidler does is ripe for commercialization, although she emphasizes that “It’s research -- to make it into a product, it would need a lot more work.”
She would be delighted to have a student take the fashion algorithm and turn it into a computer application. Meanwhile, she and Urtasun are working on an algorithm that offers suggestions on making an outfit more fashionable and displays those suggestions in the photo online “so you can see the effect.”
Another recent project involved creating an algorithm to assist a person in localizing him- or herself in a shopping mall.
“If your GPS doesn’t work and you get lost, you can snap a photo of your surroundings and it will tell you how to navigate to the store you want to visit,” Fidler says.
It’s all fascinating to a woman who says, “I was born a nerd.” After attending a talk about computer vision during university, “It was love at first sight. Everything else stopped existing.”
Fidler hopes to convey some of that passion to her UTM students.
“I really enjoy teaching,” she says. “It’s a lot of work, but a lot of fun.”