Self-Driving Cars

[Jeff Lebowski]

Yep. An attentive human probably wouldn't have avoided her, but an attentive human would have made an attempt to brake or swerve, even if it was too late. The fact that, even after she emerged from the shadows, the radar didn't pick her up is the truly disturbing part. Clearly she had no business walking across that street at that location without looking, but this doesn't seem like a complex situation that fooled the computer's algorithm. Pretty ironic set of facts for the first publicized self-driving related death, considering all the "who should the self-driving car kill" dilemmas we've seen over the last few years.

I wonder if the pedestrian had headphones on. I see that all the time and it drives me crazy. People with headphones walk out into traffic without looking either way.

[Omaha 680]

I wonder if the pedestrian had headphones on. I see that all the time and it drives me crazy. People with headphones walk out into traffic without looking either way.

In NYC they also like to be staring at their phones.

[CardiacCoug]

The psychology of fatal accidents with self-driving cars is pretty interesting.

If car accident fatalities could drop 50% tomorrow by switching completely to self-driving cars the math says we should take that immediately — would save thousands of lives.

However since car accidents are caused by substances, distractions, and stupidity we would never accept self-driving cars that were anywhere near half as bad as humans. It won’t happen until self-driving cars avoid 99%+ of accidents.

I know there are skeptics here but I think self-driving cars will happen in our lifetimes and it will be awesome. The cars will be really nervous and tentative but we will love it and think about how lame it was when you couldn’t answer emails or watch Netflix during your drive to work.

[Pelado]

The psychology of fatal accidents with self-driving cars is pretty interesting.

If car accident fatalities could drop 50% tomorrow by switching completely to self-driving cars the math says we should take that immediately — would save thousands of lives.

However since car accidents are caused by substances, distractions, and stupidity we would never accept self-driving cars that were anywhere near half as bad as humans. It won’t happen until self-driving cars avoid 99%+ of accidents.

I know there are skeptics here but I think self-driving cars will happen in our lifetimes and it will be awesome. The cars will be really nervous and tentative but we will love it and think about how lame it was when you couldn’t answer emails or watch Netflix during your drive to work.

I'm holding out for self-flying cars.

[Walter Sobchak]

The psychology of fatal accidents with self-driving cars is pretty interesting.

If car accident fatalities could drop 50% tomorrow by switching completely to self-driving cars the math says we should take that immediately — would save thousands of lives.

However since car accidents are caused by substances, distractions, and stupidity we would never accept self-driving cars that were anywhere near half as bad as humans. It won’t happen until self-driving cars avoid 99%+ of accidents.

I know there are skeptics here but I think self-driving cars will happen in our lifetimes and it will be awesome. The cars will be really nervous and tentative but we will love it and think about how lame it was when you couldn’t answer emails or watch Netflix during your drive to work.

Yeah, and if a frog had wings, it wouldn't bump its ass a-hoppin'.

Untitled.jpg

[beefytee]

https://www.wired.com/story/tesla-autopilot-investigation-ntsb-crash/

Tesla claims you are 3.7 times less likely to be involved in a fatal accident if you've got Autopilot (which it sells as a $5,000 option). “It unequivocally makes the world safer for the vehicle occupants, pedestrians and cyclists,” the company said in a recent blog post. “The consequences of the public not using Autopilot, because of an inaccurate belief that it is less safe, would be extremely severe.”

[Eddie]

https://www.wired.com/story/tesla-au...on-ntsb-crash/

When that pedestrian was hit by the self driving car a while back, I took a peek at the comments section for the article on KSL - just for entertainment value.

There was one guy who insisted that a computer would NEVER be as safe as a person, because people can look back and forth between mirrors and also check their blind spots.

I'm thinking "WHAT?!? A computer can be looking 360 degrees simultaneously. It can have eyes in the back of its head. Literally. You HONESTLY think a human can check all of the spots around the vehicle more quickly and more efficiently than a computer?"

I get that there are bugs to work out as far as what to do and how it processes the information it sees. Sure. But to say a human can see more than a computer with as many sensors as you want to attach to it? Um...no.

[Jeff Lebowski]

When that pedestrian was hit by the self driving car a while back, I took a peek at the comments section for the article on KSL - just for entertainment value.

There was one guy who insisted that a computer would NEVER be as safe as a person, because people can look back and forth between mirrors and also check their blind spots.

I'm thinking "WHAT?!? A computer can be looking 360 degrees simultaneously. It can have eyes in the back of its head. Literally. You HONESTLY think a human can check all of the spots around the vehicle more quickly and more efficiently than a computer?"

I get that there are bugs to work out as far as what to do and how it processes the information it sees. Sure. But to say a human can see more than a computer with as many sensors as you want to attach to it? Um...no.

Believe it or not, but there are lots of things humans can do and process more efficiently than computers.

[Commando]

Believe it or not, but there are lots of things humans can do and process more efficiently than computers.

Including detection of paranormal activity!

[beefytee]

Believe it or not, but there are lots of things humans can do and process more efficiently than computers.

The number of those things is shrinking by the day.

Speaking of Tesla and this:

https://www.msn.com/en-us/news/technology/tesla-relied-on-too-many-robots-to-build-the-model-3-elon-musk-says/ar-AAvQS4T?ffid=gz

Elon Musk says Tesla relied on too many robots to build the Model 3, which is partly to blame for the delays in manufacturing the crucial mass-market electric car.

[Eddie]

Believe it or not, but there are lots of things humans can do and process more efficiently than computers.

Oh, I believe it. I think it was his insistence that a computer will NEVER be able to look behind the car and check blind spots as quickly as a human that I was focused on. I have no doubt that humans process many things more quickly than a computer. Just like I have no doubt that computers process many other things more quickly than a human.

[Walter Sobchak]

Embedded Tweet

[Walter Sobchak]

As I stated earlier in this thread, AI in an open domain like the public transportation system will NEVER happen. The only way to accomplish self-driving cars is if you completely close the system akin to air traffic.

https://www.nytimes.com/2018/05/18/o...hallenges.html

A.I. Is Harder Than You Think
By Gary Marcus and Ernest Davis

Mr. Marcus is a professor of psychology and neural science. Mr. Davis is a professor of computer science.

May 18, 2018

The field of artificial intelligence doesn’t lack for ambition. In January, Google’s chief executive, Sundar Pichai, claimed in an interview that A.I. “is more profound than, I dunno, electricity or fire.”

Day-to-day developments, though, are more mundane. Last week, Mr. Pichai stood onstage in front of a cheering audience and proudly showed a video in which a new Google program, Google Duplex, made a phone call and scheduled a hair salon appointment. The program performed those tasks well enough that a human at the other end of the call didn’t suspect she was talking to a computer.

Assuming the demonstration is legitimate, that’s an impressive (if somewhat creepy) accomplishment. But Google Duplex is not the advance toward meaningful A.I. that many people seem to think.

If you read Google’s public statement about Google Duplex, you’ll discover that the initial scope of the project is surprisingly limited. It encompasses just three tasks: helping users “make restaurant reservations, schedule hair salon appointments, and get holiday hours.”

Schedule hair salon appointments? The dream of artificial intelligence was supposed to be grander than this — to help revolutionize medicine, say, or to produce trustworthy robot helpers for the home.

The reason Google Duplex is so narrow in scope isn’t that it represents a small but important first step toward such goals. The reason is that the field of A.I. doesn’t yet have a clue how to do any better.

As Google concedes, the trick to making Google Duplex work was to limit it to “closed domains,” or highly constrained types of data (like conversations about making hair salon appointments), “which are narrow enough to explore extensively.” Google Duplex can have a human-sounding conversation only “after being deeply trained in such domains.” Open-ended conversation on a wide range of topics is nowhere in sight.

The limitations of Google Duplex are not just a result of its being announced prematurely and with too much fanfare; they are also a vivid reminder that genuine A.I. is far beyond the field’s current capabilities, even at a company with perhaps the largest collection of A.I. researchers in the world, vast amounts of computing power and enormous quantities of data.

[...]

But the basic problem remains the same: No matter how much data you have and how many patterns you discern, your data will never match the creativity of human beings or the fluidity of the real world. The universe of possible sentences is too complex. There is no end to the variety of life.

[Walter Sobchak]

Had a chance to spend some time last week with a guy who does transportation research for a living. Very well-respected guy with a ton of experience and publications. One of his specialty areas right now autonomous vehicles and how they would impact road usage, traffic, etc. Fascinating stuff.

I told him I was skeptical about how quickly we would overcome the software and safety hurdles associated fully autonomous driving as it is frequently portrayed in the media and by many prognosticators. He smiled and said, "Oh, we are 40-50 years away in my opinion."

Funny how its the people who actually write software for a living that are the skeptics on autonomous driving.

Yeah, who could have predicted this?

Self-Driving Cars are Headed Toward an AI Roadblock

But the dream of a fully autonomous car may be further than we realize. There’s growing concern among AI experts that it may be years, if not decades, before self-driving systems can reliably avoid accidents. As self-trained systems grapple with the chaos of the real world, experts like NYU’s Gary Marcus are bracing for a painful recalibration in expectations, a correction sometimes called “AI winter.”

[...]

Deep learning requires massive amounts of training data to work properly, incorporating nearly every scenario the algorithm will encounter. Systems like Google Images, for instance, are great at recognizing animals as long as they have training data to show them what each animal looks like. Marcus describes this kind of task as “interpolation,” taking a survey of all the images labeled “ocelot” and deciding whether the new picture belongs in the group.

Engineers can get creative in where the data comes from and how it’s structured, but it places a hard limit on how far a given algorithm can reach. The same algorithm can’t recognize an ocelot unless it’s seen thousands of pictures of an ocelot — even if it’s seen pictures of housecats and jaguars, and knows ocelots are somewhere in between. That process, called “generalization,” requires a different set of skills.

For a long time, researchers thought they could improve generalization skills with the right algorithms, but recent research has shown that conventional deep learning is even worse at generalizing than we thought. One study found that conventional deep learning systems have a hard time even generalizing across different frames of a video, labeling the same polar bear as a baboon, mongoose, or weasel depending on minor shifts in the background. With each classification based on hundreds of factors in aggregate, even small changes to pictures can completely change the system’s judgment, something other researchers have taken advantage of in adversarial data sets.

[...]

Each accident seems like an edge case, the kind of thing engineers couldn’t be expected to predict in advance. But nearly every car accident involves some sort of unforeseen circumstance, and without the power to generalize, self-driving cars will have to confront each of these scenarios as if for the first time. The result would be a string of fluke-y accidents that don’t get less common or less dangerous as time goes on.

Have no fear though self-driving advocates, there is a solution!

Andrew Ng — a former Baidu executive, Drive.AI board member, and one of the industry’s most prominent boosters — argues the problem is less about building a perfect driving system than training bystanders to anticipate self-driving behavior. In other words, we can make roads safe for the cars instead of the other way around.

Just need to tell people to stop doing unpredictable things. Ho hum.

It's time for the self-driving car advocates to run up the white flag.

[Bo Diddley]

How long before ordinary people driving are able to learn to deal with the chaos? TIA