In previous years, calling a cab could be a bit of a pain. You needed to make sure you had cash, allowed enough time for the driver to arrive, had the number of the cheapest taxi firm, and made sure you asked for it to go on the meter so you weren’t overcharged. If you’re switched on and haven’t been living under a rock for the past few years you probably haven’t had this problem for quite some time thanks to Uber.
Uber are on a mission to create ‘possibilities for riders, drivers, and cities’, and since their launch in San Francisco in March 2009 they’re now operating in 633 cities worldwide. Since their beginnings as the app to combat taxi woes, the tech company are now a leading force in the race for safely bringing autonomous vehicles to our roads. Their Advanced Technologies Group (ATG) are currently working on self driving technologies, mapping, and vehicle safety.
Road safety has always been, and will always be, a concern in both the production of vehicles and for the general public as customers when purchasing their cars. Building the world’s safest, most reliable self-driving cars is exactly what Uber ATG are trying to do, and it’s an incredibly challenging task with constant areas for improvement. Earlier this year, Uber hired AI expert, Raquel Urtasun as they invest more time and resources in building their lab for driverless cars outside of the US, in Canada. Raquel, an associate professor at the University of Toronto, specialises in computer vision and is working to teach driverless cars to view and understand the world around them.
Raquel’s team are leveraging deep learning techniques 'to improve every facet of the technologies to optimise both efficiency and safety, and deep learning plays a key part in perception, prediction, localization, mapping and motion planning.'
At the Deep Learning Summit in Montreal this October, Raquel will be presenting her most recent breakthroughs in the field and discussing the obstacles that Uber ATG are facing to get driverless cars on the roads. Upon speaking with Raquel, she explained how she thinks that ‘self-driving technology has the potential to fundamentally change the way we live in a very positive way’ and will act as a key step towards building smarter cities that provide a better quality life where ‘AI will have a fundamental role on transforming them’. In the future, these advancements will not only improve road safety, but will provide us with cities with ‘with reduced congestion, improved mobility for those who have difficulty getting around today, greener cities, and increased access to public transit.’
Personal vehicles currently take up a huge percentage of the roads as well as parking spaces and garages, but Raquel predicts that in the future we ‘can begin to transform those parting structures into green space, new businesses, and more’, as the demand for personal vehicles decreases. Upon asking Raquel about the safety of these vehicles and how Uber ATG are using AI to create safer roads, she said:
Safety is a our number one priority at Uber ATG. Right now, we have well-trained safety drivers operating our fleet, and we constantly learn from their time on the road. In particular, the operator in the passenger seat documents important interactions with the vehicle that we, as engineers, use to improve our software on a daily basis. This iterative process allows us to identify areas where AI can improve the rider experience that ensures a safe ride for those inside the vehicle, and outside. We use deep learning to make the technology more reliable every single day.
Raquel Urtasun is the Head of Uber ATG Toronto. She is also an Associate Professor in the Department of Computer Science at the University of Toronto, a Canada Research Chair in Machine Learning and Computer Vision and a co-founder of the Vector Institute for AI. To learn more from Raquel, register for the Deep Learning Summit, Montreal, October 10 & 11.
Interested in learning more about autonomous vehicles? Watch the panel discussion 'How can we apply ML and DL to accelerate the autonomous vehicle?' from the Machine Intelligence in Autonomous Vehicle Summit earlier this summer.