U of T students win first place in AutoDrive Challenge for engineering self-driving car

Victory marks second consecutive year that aUToronto came first in the international competition for car named Zeus

U of T students win first place in AutoDrive Challenge for engineering self-driving car

A team of U of T engineering and computer science students won first place in an international self-driving car competition in Michigan in July. The group, named aUToronto, pitted its vehicle Zeus against those from seven other North American universities.

This victory marks the second consecutive year that the team has won first place in the AutoDrive Challenge, as it did last year against the same competitors. The competition, run by the Society of Automotive Engineers (SAE) International and General Motors, will hold its final round for this three-year cycle in 2020.

The team scored first in eight categories

aUToronto won first place in eight of nine categories this year, defeating competitors from the University of Waterloo, Michigan State University, Michigan Tech University, Kettering University, Virginia Tech, North Carolina A&T State University, and Texas A&M University.

For its ability to recognize traffic signs, such as speed limits and ‘do not enter’ signs, as well as respond appropriately to them, aUToronto’s Zeus won first place in the Traffic Control Sign Challenge.

aUToronto’s car also placed first in the Pedestrian Challenge, which tested cars on their ability to wait for pedestrian replicas to completely cross a road before proceeding, as well as in the MCity Challenge, which required the vehicles to navigate around obstacles such as a tunnel and railroad crossing.

“Correctly detecting and classifying all the traffic lights and signs was more difficult than we anticipated,” reflected aUToronto Team Lead Keenan Burnett in an email to The Varsity.

The team’s approach to the problem, which utilized deep neural networks, or systems of artificial neurons, required substantial tuning and data collection to work effectively.

Zeus further secured first place in the Intersection Challenge and tied for first in the Mapping Challenge.

One key aspect to the team’s success here can be attributed to the realization that “relying only on GPS/IMU for positioning [could have been] risky.”

“We opted to integrate a more advanced localization software that uses [a laser system called] LIDAR instead,” wrote Burnett. “This proved to be one [of] the keys to our success at the competition.”

aUToronto also scored first in the categories of Social Responsibility Report, Social Responsibility Presentation, and Concept Design Presentation. The team placed second in Concept Design.

Origins of aUToronto and Zeus

“The team’s inception traces back to when SAE was soliciting applications from universities to compete in their new self-driving competition,” wrote Burnett.

“The idea is that they would select the [eight] top university applications based on the quality of the proposals, the backing of the university, and facilities that would be made available to students.”

Cristina Amon, U of T’s Dean of Engineering at the time, requested Professor Tim Barfoot to submit a proposal in 2017, according to Burnett. Professors Barfoot and Angela Schoellig, from the U of T Institute for Aerospace Studies (UTIAS), partnered to submit the proposal and were selected as one of the eight competing university teams.

Burnett, who was then applying to be a graduate student at UTIAS, asked Barfoot if a position was available to run the team. He was hired after an interview in April.

“From then, I built the year 1 team up from scratch,” wrote Burnett. “We hired a small set of students with excellent technical and leadership skills to head the sub-teams for the first year of the competition.”

Burnett then let the sub-team leads hire their own sub-team members. New members have been recruited roughly every four months since the fall of 2017.

Around 100 students are on the team, according to Burnett. Undergraduates comprise 90 per cent of the team, while graduate students make up the remaining 10 per cent. Students primarily study electrical engineering, mechanical engineering, engineering science, and computer science.

In terms of resources, aUToronto received a Chevrolet Bolt Electric Vehicle and an Intel compute server from the competition’s organizers. The team then acquired sensors and developed infrastructure around the Bolt to turn it into a fully self-driving car.

“The turnaround between receiving the vehicle and shipping it off to compete in the Year 1 competition was just 6 months,” wrote Burnett.

The team then competed in the second round of the AutoDrive Challenge earlier this year.

“We still have a third year of the competition coming up,” wrote Burnett. “It will be held at the Ohio Transportation Research Center. We anticipate needing to handle dynamic [challenges] and [drive] at much higher speeds.”

U of T undergraduate co-wins prestigious research award at AIES Conference

Inioluwa Deborah Raji awarded best paper for detecting facial recognition bias in Amazon technology

U of T undergraduate co-wins prestigious research award at AIES Conference

Amazon’s facial recognition technology may be misidentifying dark-skinned women, according to U of T Engineering Science undergraduate Inioluwa Deborah Raji and Massachusetts Institute of Technology Media Lab research assistant Joy Buolamwini. This finding helped Raji and Buolamwini win “best student paper” at the Artificial Intelligence, Ethics, and Society (AIES) Conference in Honolulu, Hawaii. Held in January, the prestigious conference was sponsored by Google, Facebook, Amazon, and the like.

Their paper, which caught the Toronto Star’s attention, was a follow-up on an earlier audit by Buolamwini on technology from Microsoft, IBM, and Face++, a facial recognition startup based in China.

Origins of the research

Buolamwini’s earlier study, “Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification,” investigated the accuracy of artificial intelligence (AI) systems used by the three technology firms for facial recognition. Then-Microsoft Research computer scientist Timnit Gebru co-authored the paper.

Raji wrote that after reading about Buolamwini’s experiences “as a black woman being regularly misgendered by these models,” she wondered if her personal experience would hold true for a larger dataset containing samples of other dark-skinned women. This proved to be the case in the final analysis.

According to Raji, “Gender Shades” uncovered “serious performance disparities” in software systems used by the three firms. The results showed that the software misidentified darker-skinned women far more frequently than lighter-skinned men.

In an email to The Varsity, Raji wrote that since the release of Buolamwini and Gebru’s study, all three audited firms have updated their software to address these concerns.

For the paper submitted to the AIES Conference, Raji and Buolamwini tested the updated software to examine the extent of the change. They also audited Amazon and Karios, a small technology startup, to see how the companies’ adjusted performance “compared to the performance of companies not initially targeted by the initial study.”

At the time of Raji and Buolamwini’s follow-up study in July, Raji wrote that “the ACLU [American Civil Liberties Union] had recently reported that Amazon’s technology was being used by police departments in sensitive contexts.”

Amazon denied that bias was an issue, saying that it should not be a concern for their “partners, clients, or the public.”

Raji and Buolamwini’s study showed evidence to the contrary. “We found that they actually had quite a large performance disparity between darker females and lighter males, not working equally for all the different intersectional subgroups,” wrote Raji.

Amazon’s response to the study

In a statement sent by Amazon’s Press Center to The Varsity, a representative wrote that the results of Raji and Boulamwini’s study would not be applicable to technologies used by law enforcement.

Amazon wrote that the study’s results “are based on facial analysis and not facial recognition,” and clarified that “analysis can spot faces in videos or images and assign generic attributes such as wearing glasses,” while “recognition is a different technique by which an individual face is matched to faces in videos and images.”

“It’s not possible to draw a conclusion on the accuracy of facial recognition for any use case – including law enforcement – based on results obtained using facial analysis,” continued Amazon. “The results in the paper also do not use the latest version of Rekognition and do not represent how a customer would use the service today.”

In a self-study using an “up-to-date version of Amazon Rekognition with similar data downloaded from parliamentary websites and the Megaface dataset of 1M images,” explained Amazon, “we found exactly zero false positive matches with the recommended 99% confidence threshold.”

However, Amazon noted that it continues “to seek input and feedback to constantly improve this technology, and support the creation of third party evaluations, datasets, and benchmarks.” Furthermore, Amazon is “grateful to customers and academics who contribute to improving these technologies.”

The pair’s research could inform policy

Raji wrote that while it’s tempting for the media to focus on the flaw in Amazon’s software, she thinks that the major contribution of her paper is in helping to uncover how researchers can effectively conduct and present an audit of an algorithmic software system to prompt corporate action.

“Gender Shades introduced the idea of a model-level audit target, a user-presentative test set, a method for releasing results to companies called Coordinated Bias Disclosure,” wrote Raji.

In other words, Raji and Buolamwini’s research showed an effective way for companies and policymakers to investigate and communicate a problem in software systems and take action.

Most importantly, wrote Raji, the studies highlight the need for researchers to evaluate similar software models “with an intersectional breakdown of the population being served.”