Uber self-driving SUV fatal accident – a Computer Scientist’s views

Thursday, 22 March, 2018

20180324 update: For now, I’ve found these two posts by Brad Templeton to be very insightful and cover some of the issues that I want to write about but Brad wrote in much more detail! Have a read, 03/20 “New facts and questions on Uber robocar fatality” & 03/21 “It certainly looks bad for Uber“. I may still add more if I see more facts of the case especially when Uber starts to voluntarily (or be compelled to) provide more of its internal technical data. I hope Uber won’t try to brush this fatality under the carpet. Will see.

***

I just read some news reports and watched the video of the Uber self-driving SUV fatal accident. (WARNING: Video contains disturbing images. Viewer discretion is advised.) I know I do not have full information yet so I hope to share my views (for now, semi-technical/semi-informed) on this Uber self-driving fatal accident as best as I can. And in the coming days when I have time, I hope to keep updating this post when more technical and police investigative information are available.

A bit of background first. In 2013 February (more than 5 years ago now), I was already interested in driverless technologies and already interviewed U of T Professor Emeritus C.C. Kelly Gotlieb, “Father of Computing in Canada”, to talk about many topics including Google driverless car and issues like whose to blame when an accident happened? Sadly, we now have a fatal accident on hand to talk about.

From the AP report “Experts: Uber self-driving system should have spotted woman”, this Uber self-driving SUV is using LIDAR laser sensors technology to “see”. (note: LIDAR stands for Light Detection and Ranging and it “measures distance to a target by illuminating the target with pulsed laser light” which can see perfectly well even in total darkness as it uses laser.) I made this observation re LIDAR in direct response to this sentence of the news report, “The lights on the SUV didn’t illuminate 49-year-old Elaine Herzberg on Sunday night until a second or two before impact, raising questions about whether the vehicle could have stopped in time.” And the fact the Uber safety driver was NOT paying attention to the road when he killed the 49-year-old Elaine Herzberg!

Let me quote from the AP report “Experts: Uber self-driving system should have spotted woman”,

““The victim did not come out of nowhere. She’s moving on a dark road, but it’s an open road, so Lidar (laser) and radar should have detected and classified her” as a human, said Bryant Walker Smith, a University of South Carolina law professor who studies autonomous vehicles.

Smith said the video may not show the complete picture, but “this is strongly suggestive of multiple failures of Uber and its system, its automated system, and its safety driver.”

Sam Abuelsmaid, an analyst for Navigant Research who also follows autonomous vehicles, said laser and radar systems can see in the dark much better than humans or cameras and that Herzberg was well within the range.

“It absolutely should have been able to pick her up,” he said. “From what I see in the video it sure looks like the car is at fault, not the pedestrian.”

Smith said that from what he observed on the video, the Uber driver appears to be relying too much on the self-driving system by not looking up at the road.

“The safety driver is clearly relying on the fact that the car is driving itself. It’s the old adage that if everyone is responsible no one is responsible,” Smith said. “This is everything gone wrong that these systems, if responsibly implemented, are supposed to prevent.”

The experts were unsure if the test vehicle was equipped with a video monitor that the backup driver may have been viewing.

Uber immediately suspended all road-testing of such autos in the Phoenix area, Pittsburgh, San Francisco and Toronto. The National Transportation Safety Board, which makes recommendations for preventing crashes, is investigating the crash.”

I will try to come back to this article and add more details and updates in the coming days when I have more time. Will see.

For now, here is the particular segment of my 5 years old 2013 interview with Prof. Gotlieb talking about “Google [and by extension, any other company’s] Driverless Car gets into an accident, whose to blame? And who can you sue? The person who wrote the program? Google who authorize the car? Car manufacture? The person who is in the car? Or all of the above? […] Lots of questions to be asked when failure happen.”

xxx

Advertisements

Cybersecurity of Voting Machines

Tuesday, 5 December, 2017
Dr. Matt Blaze's House testimony on the security of voting machines.

Dr. Matt Blaze’s House testimony on the security of voting machines.

#VotingMachines #eVoting It worries me that some form of e-voting was used in last Calgary municipal elections and more are being studied to be potentially used in the future. (Case of I don’t know enough.) As someone who has been following e-voting and development of secure voting machines for decades (a company I used to work for had a team that develop e-voting system), I have my serious reservations with e-voting and voting machines and want all levels of Canadian governments (city, provincial, federal) to study slow and proceed very very very carefully!

To learn more, I’m watching UPenn’s Dr. Matt Blaze‘s House testimony on the security of voting machines.

Cybersecurity of Voting Machines (26m45s)

More of Dr. Blaze‘s testimonies here at these timecodes: 35m30s ; 54m19s ; 1h5m56s ; 1h30m30s ; 1h44m02s ; 1h48m25s and following individually video links to specific timecode segments.

Read the rest of this entry »


Meet Geoffrey Hinton, U of T’s Godfather of Deep Learning

Thursday, 26 October, 2017

//Meet Geoffrey Hinton: U of T Professor Emeritus of computer science, an Engineering Fellow at Google, and Chief Scientific Adviser at the Vector Institute for Artificial Intelligence.

In this interview with U of T News, Prof. Hinton discusses his career, the field of artificial intelligence and the importance of funding curiosity-driven scientific research.//

Proud to be a UT computer science grad. Wish I had taken a class from Prof. Hinton.

Meet Geoffrey Hinton, U of T’s Godfather of Deep Learning

This video of “Panel of Pioneers” at RE-WORK (Deep Learning Summit Track 1, Montreal, 2017) is a great watch.

Intro from Yann (FB Director of AI Research) LeCun’s FB page.

//Video of the panel in which Yoshua Bengio, Geoffrey Hinton and I answer questions moderated by Joelle Pineau (who leads the FAIR-Montréal lab).

This took place at the Re*Work deep learning summit in Montreal a few weeks ago.//

Intro from the web page (with video, ~23 minutes).

//Overview

The Deep Learning Summit took place in Montreal on 10-11 October 2017 and brought together global AI pioneers including: Yoshua Bengio, Yann LeCun, and Geoffrey Hinton, as well as experts from companies including Intel, NVIDIA, Twenty Billion Neurons and Apple.

We’re currently working on the videos for the summit so please fill in the form below and we’ll email you when they’re ready.//


U of Toronto engineering researchers mend broken hearts with expanding tissue bandage

Friday, 25 August, 2017

Screen Shot 2017-08-25 at 9.56.43 AM - New biomaterial developed by U of T engineering researchers could be delivered through minimally invasive surgery

Very cool news. Excerpts from University of Toronto news “New biomaterial developed by U of T engineering researchers could be delivered through minimally invasive surgery” (emphasis, extra note & links added) (for an in-depth look, see technical article, Nature Materials “Flexible shape-memory scaffold for minimally invasive delivery of functional tissues” ),

A team of U of T engineering researchers is mending broken hearts with an expanding tissue bandage a little smaller than a postage stamp.

Repairing heart tissue destroyed by a heart attack or medical condition with regenerative cells or tissues usually requires invasive open-heart surgery. But now biomedical engineering Professor Milica Radisic [K’s note: including links to PubMed listed articles] and her colleagues have developed a technique that lets them use a small needle to inject a repair patch, without the need to open up the chest cavity.

Radisic’s team are experts in using polymer scaffolds to grow realistic 3D slices of human tissue in the lab. One of their creations, AngioChip, is a tiny patch of heart tissue with its own blood vessels – the heart cells even beat with a regular rhythm. Another one of their innovations snaps together like sheets of Velcro™.

Such lab-grown tissues are already being used to test potential drug candidates for side-effects, but the long-term goal is to implant them back into the body to repair damage.

“If an implant requires open-heart surgery, it’s not going to be widely available to patients,” says Radisic.

She says that after a myocardial infarction – a heart attack – the heart’s function is reduced so much that invasive procedures like open-heart surgery usually pose more risks than potential benefits.

“It’s just too dangerous,” she says.

Miles Montgomery, a PhD candidate in Radisic’s lab, has spent nearly three years developing a patch that could be injected, rather than implanted. [K’s note: more news on Miles]

“At the beginning, it was a real challenge,” he says. “There was no template to base my design on, and nothing I tried was working. But I took these failures as an indication that I was working on a problem worth solving.”

After dozens of attempts, Montgomery found a design that matched the mechanical properties of the target tissue and had the required shape-memory behaviour: as it emerges from the needle, the patch unfolds itself into a bandage-like shape.

[…]

The scaffold is built out of the same biocompatible, biodegradable polymer used in the team’s previous creations. Over time, the scaffold will naturally break down, leaving behind the new tissue.

The team also showed that injecting the patch into rat hearts can improve cardiac function after a heart attack: damaged ventricles pumped more blood than they did without the patch.

“It can’t restore the heart back to full health, but if it could be done in a human, we think it would significantly improve quality of life,” says Radisic.

There is still a long way to go before the material is ready for clinical trials. Radisic and her team are collaborating with researchers at the Hospital for Sick Children to assess the long-term stability of the patches, as well as whether the improved cardiac function can be maintained.

They have also applied for patents on the invention and are exploring the use of the patch in other organs, such as the liver.

“You could customize this platform, adding growth factors or other drugs that would encourage tissue regeneration,” says Radisic. “I think this is one of the coolest things we’ve done.”

Injectable tissue patch could help repair damaged organs – U of T Engineering


Julie Payette – Canada’s next Governor General

Thursday, 13 July, 2017

I’m thrilled and excited to hear Ms. Julie Payette, TA of myUniversity of Toronto Computer Science CSC258 class (I wrote more in this post), has been named Canada’s next Governor General.

2017 July 13, CBC News, “‘Unquestionably qualified’: Ex-astronaut Julie Payette formally introduced as Canada’s next GG – Prime minister holds news conference on Parliament Hill to name successor to David Johnston

U of T News, “U of T alumna Julie Payette to be next Governor General

Via CBC Politics LIVE FB post.

Have a watch of this amazing CBC Witness (1993) documentary “Space For Four (1993)


The Usefulness of Useless Knowledge

Tuesday, 30 May, 2017

I’m watching this great talk thanks to Yann LeCun’s FB post. I’m also planning to read “The Usefulness of Useless Knowledge” by Abraham Flexner (PDF via IAS). Fascinating stuff.

Robbert Dijkgraaf: “The Usefulness of Useless Knowledge” | Talks at Google


The Future of Go Summit: Ke Jie & AlphaGo

Tuesday, 23 May, 2017

Master” is the new version of “AlphaGo” which Demis Hassabis stated, in the post game press conference with 9 dan Go player Ke Jie (柯潔), the details will be published for others to study similar to AlphaGo’s Nature article.

Wired, “An Improved AlphaGo Wins Its First Game Against the World’s Top Go Player

Last year, in South Korea, AlphaGo topped the Korean grandmaster Lee Sedol, becoming the first machine to beat a professional Go player—a feat that most AI researchers believed was still years away, given the extreme complexityof the ancient Eastern game. Now, AlphaGo is challenging Ke Jie, the current world number one.

According to Demis Hassabis, the CEO and founder of DeepMind, this time out the machine is driven by a new and more powerful architecture. It can now learn the game almost entirely from play against itself, relying less on data generated by humans. In theory, this means DeepMind’s technology can more easily learn any task.

MIT Technology review, “Intelligent Machines A Stronger AlphaGo Defeats the World’s Number One Player

The Future of Go Summit, Match One: Ke Jie & AlphaGo

May 26, 2017 Update:

Wired, “Google’s AlphaGo Trounces Humans—But It Also Gives Them a Boost

Much of that future has yet to play out. And there is no guarantee that AI improves humanity. “In some cases,” grandmaster Gu Li said after a pair game alongside AlphaGo, “I could not follow in his footsteps.” But certainly, DeepMind has effected real change in the world of Go, a game that’s enormously popular across China, Korea, and other parts of Asia, and that is a comforting thing. In at least one way, AI has helped make humans better.

After losing matches to AlphaGo, European champion Fan Hui and Korean grandmaster Lee Sedol said the machine opened their eyes to new possibilities. This raised awareness was on wide display this week in China, when Ke Jie opened the first game with a strategy straight from the AlphaGo playbook.

Ke Jie went on to lose that game and then the next. And some observers continued to lament that machines were eclipsing humans. But that’s not the story of AlphaGo’s trip to China. What’s most striking is how closely the players have studied the games played by AlphaGo—and how hungry they are for more. Many have repeatedly called on DeepMind to release the many games that AlphaGo has played in private. They know they can’t beat the machine. But like Thore Graepel, they believe it can make them better.

The Future of Go Summit, Match Two: Ke Jie & AlphaGo Read the rest of this entry »


%d bloggers like this: