Vector

After years of conducting world-leading research on self-driving vehicles, U of T Associate Professor Raquel Urtasun is joining Uber – and bringing the American company’s first international research lab to Toronto.

“I’m proud to welcome Raquel Urtasun to Uber,” said Uber CEO Travis Kalanick in a blog posting announcing the deal today.

“Raquel will remain in Toronto to lead a new branch of our Advanced Technologies Group – our first outside the U.S.,” he continued, adding the company hopes “to draw from the region’s impressive talent pool as we grow.”

Kalanick added that Urtasun and her team “will further strengthen our self-driving engineering efforts in San Francisco and Pittsburgh. And, their work will complement the research underway at Uber AI Labs, led by Zoubin Ghahramani – a proud �ؿ�ζSM alum himself.”

U of T's Deep Genomics applies AI to accelerate drug development for genetic conditions

ullahnor — Wed, 03 May 2017 21:10:58 +0000

U of T's Deep Genomics applies AI to accelerate drug development for genetic conditions

ullahnor Wed, 05/03/2017 - 17:10

Cutline

U of T Engineering Professor Brendan Frey is the founder and CEO of Deep Genomics, a startup company applying deep-learning techniques to revolutionize genomic medicine (photo courtesy of Deep Genomics)

Marit Mitchell

Author legacy

Marit Mitchell

Topic

Global Lens

Artificial Intelligence

Vector

Brendan Frey

Entrepreneurship

Faculty of Applied Science & Engineering

Startup

Drugs

Faculty of Arts & Science

Donnelly

Faculty of Medicine

Subheadline

U of T spinoff company combines leading research in both machine learning and genomic science to accelerate development of highly tailored medical treatments

Genetic mutations are the cause of countless diseases and disorders, from cancer to autism to cystic fibrosis.

Now, startup company Deep Genomics is applying decades of research into machine learning and genomic science to develop genetic medicines – accelerating treatments that address the root causes of these conditions.

“If you have smoke billowing out of the tailpipe of your car, you don’t just put a filter on the tailpipe – you have to look under the hood and address the original problem,” says Brendan Frey, the co-founder and CEO of Deep Genomics, and a U of T engineering professor with cross-appointments in the department of computer science and the Donnelly Centre for Cellular and Biomolecular Research. “That’s what we’re doing: applying our platform for the discovery-phase development of medicines that address genetic problems.”

Developing new drugs is expensive, slow and inefficient – when researchers identify a protein involved in a disease, pharmaceutical companies often use a ‘guess-and-test’ approach to see whether any of the known drug molecules in their arsenal is a match to the protein’s unique shape. Often, thousands of molecules need to be screened in order to generate a match.

Frey’s team at Deep Genomics is looking at the first biological step in the process: at the genes that contain the blueprints for proteins and instructions on how and when to produce them.

“There are many ways a protein could be causing a problem, resulting from different changes to the genome. We can see those changes at the level of individual genes,” says Frey. “Instead of focusing on proteins, we’re focusing on the genetic mutations that are the source of the problem.”

Learn more about entrepreneurship and startups at U of T

Tracking proteins using AI: U of T scientists develop deep learning algorithm

ullahnor — Wed, 03 May 2017 15:56:25 +0000

Tracking proteins using AI: U of T scientists develop deep learning algorithm

ullahnor Wed, 05/03/2017 - 11:56

Cutline

Yeast cells (purple) with DNA-containing nuclei (pink) and protein (green) residing in a cell's waste compartment

Jovana Drinjakovic

Author legacy

Jovana Drinjakovic

Topic

Global Lens

Deep Learning

Artificial Intelligence

Faculty of Applied Science & Engineering

Faculty of Medicine

machine learning

U of T researchers have developed a deep learning algorithm that can track proteins to help reveal what makes cells healthy and what goes wrong in disease.

From self-driving cars to computers that can diagnose cancer, artificial intelligence (AI) is shaping the world in ways that are hard to predict, but for cell biologists, the change could not have come soon enough. With new and fully automated microscopes, scientists collect reams of data faster than they can analyze it.

The protein-tracking algorithm, Dubbed DeepLoc, can recognize patterns in the cell made by proteins better and much faster than the human eye or previous computer vision-based approaches.

“We can learn so much by looking at images of cells. How does the protein look under normal conditions? Do they look different in cells that carry genetic mutations or when we expose cells to drugs or other chemical reagents?” says Benjamin Grys, a graduate student in molecular genetics who recently co-authored a paper on the research. “People have tried to manually assess what’s going on with their data, but that takes a lot of time.”

In the cover story of the latest issue of Molecular Systems Biology, teams led by the Donnelly Centre's Brenda Andrews and Charles Boone, both professors of molecular genetics, also describe DeepLoc’s ability to process images from other labs, illustrating its potential for wider use.

“Right now, it only takes days to weeks to acquire images of cells and months to years to analyze them. Deep learning will ultimately bring the timescale of this analysis down to the same timescale as the experiments,” says Oren Kraus, a lead co-author on the paper and a graduate student co-supervised by Andrews and the Donnelly Centre's Brendan Frey, professor of electrical and computer engineering.

Learn more about Frey's startup which uses AI to develop drugs for genetic conditions

Kraus is now working with Jimmy Ba, a graduate student of AI pioneer Geoffrey Hinton, a U of T University Professor Emeritus in computer science who is the chief scientific adviser of the newly established Vector Institute, to commercialize the method through a new startup. The goal of the startup named, Phenomic AI, is to analyse cell image-based data for pharmaceutical companies.

“In an image based drug screen, you can actually figure out how the drugs are affecting different cells based on how they look rather than some simplified parameters such as live/dead or cell size,” says Kraus. “This way you can extract a lot more information about cell state form these screens. We hope to make the early drug discovery process all the more accurate by finding more subtle effects of chemical compounds.”

Similar to other types of AI, in which computers learn to recognize patterns in data, DeepLoc was trained to recognize diverse shapes made by glowing proteins – labelled a fluorescent tag that makes them visible – in cells. But unlike computer vision that requires detailed instructions, DeepLoc learns directly from image pixel data, making it more accurate and faster.

Grys and Kraus trained DeepLoc on previously published data that shows an area in the cell occupied by more than 4,000 yeast proteins – three-quarters of all proteins in yeast. This dataset remains the most complete map showing exact position for a vast majority of proteins in any cell. When it was first released in 2015, the analysis was done with a complex computer vision and machine learning pipeline that took months to complete. DeepLoc crunched the data in a matter of hours.

DeepLoc was able to spot subtle differences between similar images. The initial analysis identified 15 different classes of proteins, each representing distinct neighbourhoods in the cell. DeepLoc identified 22 classes. It was also able to sort cells whose shape changed due to a hormone treatment, a task that the previous pipeline couldn’t complete.

Grys and Kraus were able to quickly retrain DeepLoc with images that differed from the original training set, showing that it can be used to process data from other labs. They hope that others in the field – where looking at images by eye is still the norm – will adopt their method.

“Someone with some coding experience could implement our method,” says Grys. “All they would have to do is feed in the image-training set that we’ve provided and supplement this with their own data. It takes only an hour or less to retrain DeepLoc and then begin your analysis.”

Learn more about entrepreneurship and startups at U of T

Toronto's Vector Institute officially launched

ullahnor — Thu, 30 Mar 2017 22:37:17 +0000

Toronto's Vector Institute officially launched

ullahnor Thu, 03/30/2017 - 18:37

Cutline

U of T Associate Professor Raquel Urtasun (second from right) with (from left) Minister of Finance Bill Morneau, Ontario Minister of Research, Innovation and Science Reza Moridi, Premier Kathleen Wynne, Mayor John Tory and U of T President Meric Gertler

Jennifer Robinson

Author legacy

Jennifer Robinson

Topic

Breaking Research

Vector

Artificial Intelligence

Subheadline

Aims to produce the world’s largest number of deep learning graduates

With the creation of the Toronto-based Vector Institute, Ontario and Canada are choosing to lead in the booming field of artificial intelligence, said Ontario Premier Kathleen Wynne today.

“We can own this space. This is who we are,” she told a crowded room filled with �ؿ�ζSM researchers, officials from all three levels of government and top Canadian companies who are supporting Vector.

The independent institute aims to produce more deep learning grads than any other institution in the world – as part of an effort to produce, attract and retain top talent. It will also build on the existing expertise of the globally renowned deep learning team at U of T.

The applications and implications of this ground-breaking research could be seen in demonstrations on self-driving vehicle technology by Raquel Urtasun, a U of T computer science associate professor, as well as U of T startup Deep Genomics.

Founded by Professor Brendan Frey of the Edward S. Rogers Sr. Department of Electrical & Computer Engineering, Deep Genomics uses deep learning to predict the molecular effects of genetic variation.

Read about the Vector Institute

Sanja Fidler (right), an assistant professor of computer science at U of T, focuses on object recognition, 3D scene understanding, and combining vision and language. She explains her research at Thursday's event to (from left to right) Minister of Finance Bill Morneau, Premier Kathleen Wynne and Mayor John Tory (photo by Lisa Lightbourn)

In addition to $50 million in provincial support, Vector will also receive $40 million to $50 million as part of the Government of Canada’s Pan-Canadian Artificial Intelligence Strategy, federal Finance Minister Bill Morneau said. The $125-million strategy is also supporting similar institutes in Montreal and Edmonton.

“We are home to some of the top talent when it comes to artificial intelligence,” Canadian Prime Minister Justin Trudeau said at an event earlier in the day in Brampton. “We can't afford to lose that competitive advantage, and all the good jobs that come along with it.”

More than 30 companies have also committed a combined total of more than $80 million over 10 years to support the Vector Institute, reflecting the transformational potential of deep learning and machine learning in fields as diverse as health care, finance, insurance, education, retail, advanced manufacturing, construction and transportation.

“What an amazing moment,” said U of T President Meric Gertler. “The governments of Ontario and Canada and our industry partners deserve huge credit for the visionary leadership represented in today’s announcement. On behalf of the �ؿ�ζSM, and our sister universities: Thank you for enabling us to seize this incredibly exciting opportunity.”

U of T President Meric Gertler talked about how the university has long been considered a global leader in artificial intelligence research, and that expertise can now act as an anchor to bring together researchers, government and private sector actors through the Vector Institute (photo by Lisa Lightbourn)

Using his time at the podium, Geoffrey Hinton, a U of T University Professor Emeritus in computer science and Vector’s new chief scientific adviser, deviated from the usual prepared remarks to deliver a pithy TED Talk-like PowerPoint on how neural networks and deep learning work.

Toronto Mayor John Tory said the city’s diversity and livability will help attract the talent Vector and the AI industry in Canada needs to grow and prosper. The Vector Institute is “one more thing putting us on the map,” he said. “I want us to be leaders. I want us to compete. I want us to be the home of disruptors!”

U of T University Professor Emeritus Geoffrey Hinton, who is vice-president engineering fellow at Google and will serve as the chief scientific adviser of the newly created Vector Institute, gave the audience a mini-lecture about artificial intelligence (photo by Lisa Lightbourn)

Vector

U of T's self-driving vehicle superstar to lead Uber's new research lab in Toronto

Read more at the Toronto Star

Read more at Bloomberg

Read more about the Vector Institute

Read more about Urtasun's new role at Uber

U of T's Deep Genomics applies AI to accelerate drug development for genetic conditions

Read more about startup by Frey's student using AI to analyze protein data

Learn more about entrepreneurship and startups at U of T

Tracking proteins using AI: U of T scientists develop deep learning algorithm

Learn more about Frey's startup which uses AI to develop drugs for genetic conditions

Learn more about entrepreneurship and startups at U of T

Toronto's Vector Institute officially launched

Read about the Vector Institute