Uber driverless car in fatal accident

[Fastpedeller]

kev1974 said:

The only place I have ever been where I found the locals to absolutely refuse to cross the road away from proper crossings was Poland. Spent a few months working there and the guys were totally paranoid about being done for jaywalking.

When in Italy 20 years ago we were told that the pedestian crossing (similar to uk) were there for the peds to cross the road- but the traffic would keep moving! It was something I found difficult to adapt to, although my Wife coped with it easier. I actually felt safer crossing the road in Italy, and the traffic did, indeed, miss us. There was quite heavy traffic including lots of mopeds but they did appear to be more alert and able. On returning to the UK, I was (correctly) going across a zebra crossing, and half way across looking around (as I do just in case), when I narrowly missed being run over by a white van that made no attempt to stop - I leapt out of the way in time, alas I didn't whack side of his van, such was my priority to stay alive.

[Vanden Saab]

98elise said:

Conversely we are much more likely to experience a car suddenly pulling out, or a pedestrian stepping into the road. Humans are pretty crap at dealing with situations like that.

Autonomous cars will not deal with every unusual situation, but they will deal with the majority of usual and predictable situations better than us.

They just have to be better than us to make sense, they don't have to be perfect. Even if they could just tackle motorway driving then I would pay extra for one.

So in 25 pages we have gone from "autonomous cars will save millions of lives" to "they only have to be better than us" and at the same time it is suggested that these new wonder cars should have limited liability in the event of a death or serious injury. It is an insurance companies wet dream......and the naysayers are rubbished for even daring to question it being a good idea........

[Einion Yrth]

akirk said:

ash73 said:

akirk said:

ash73 said:

akirk said:

Jonesy23 said:

That isn't AI. That's (mostly) Computer Vision. Interesting stuff albeit pretty simple to do these days; OpenCV does almost anything you'd ever want. Sign recognition used to be hard, these days it's an afternoon demo.

Mostly when people talk 'AI' they don't really mean AI. And even if they did mean to use AI in autonomous cars they need to understand what specifically they mean and whether it's a proper solution or just technical wkery because AI is a popular concept right now.

exactly - there is so much guff talked in this area - AI is artifical intelligence - i.e. not from a living being
there is no intelligence in any of this - it is simply following pre-programmed logic - if A -> B

Any system that can learn and take actions to maximise its chances of achieving a goal is AI. It doesn't need to be conscious to be AI. For example a neural net can learn new patterns that weren't directly programmed; if it detects a new pattern it might ask what the object is, or test all the response outcomes itself by scoring them, then it will recognise the object in the future and know what to do. This is also how humans learn.

In fact it has to be NOT conscious to be AI
Learning of new patterns though is not real AI - real AI is the ability to jump logic / to derive logic / to create new logic - adding a new pattern to your data bank and evaluating it against knowns, and then scoring based on a known algorithm is not AI... reality is that we know so little about how the human brain works we are a very long way from being able to replicate even a small part of it... however for Autonomous cars, the majority of what is needed is not AI, so it is not an issue - we don't yet need the car to create original poetry to read to its occupants yet!

I disagree. The coding or logic of a neural net algorithm doesn't need to be modified to learn, it learns by recognising and storing new data patterns and identifying the best actions by scoring the outcomes (or by asking its programmer!). As it learns it becomes better at making strategic choices and exceeds the sum of its parts.

Also, if a sufficiently advanced AI became conscious, and even reprogramed itself, then it would still be artificial intelligence because it's a machine, although it might be afforded the same rights as humans.

Your definition is too narrow, it's like saying life isn't life unless it can do calculus.

We will have to disagree - all that a machine based logic can do sits within the original coding - all it can 'learn' is to make choices based upon what it stores in its data bank - it can never break out of that - real intelligence has non-rational jumps and changes which come from nowhere other than the 'genius' of that brain - current machine logic can not write something creative - it can randomly assemble words and it can even analyse other examples of books to determine what is popular - but it can never create something new - something not based on the logic it holds - other than through randomisation - it can never be a Milton, or Bacon, an Austen of Dickens...

So, current levels of machine 'AI' are assemblers and organisers / prioritisers and logicians - they are never creators...

So, potentially wrong, it's almost funny; do some research.

[akirk]

Einion Yrth said:

So, potentially wrong, it's almost funny; do some research.

surely something is right or wrong, potentially wrong must make it right?!
sorry, but I have worked in IT and related worlds for 30+ years, and we are way off true AI...
conceptually it is impossible for a non-sentinent being to ever have true intelligence, unless something is conscious it can not create with personal meaning, its actions and conclusions are in a vaccuum - we can look at them externally and derive meaning because we apply our own context and beliefs, but a machine is and always will be totally incapable of approximating human intelligence, so while machines will be capable of deriving phenomenally complex conclusions, beyond the comprehension of humans, and so we will label them as AI, they can never breathe / think / philosophise / create as a human - until we gain God's skills in creation, machines will just be fast number crunchers, that does not make them true AI - the sci-fi novel fantasy of machines that can a tually think will never happen - let us not confuse complexity with originality...

anyhow - now seriously off track for this topic, the 'AI' in an AV is simply not intelligent, it uses case studies and derives logic against pre-coded formulas to try and answer questions around go / stop - they are fundamentally very basic, and very flawed, there is no way they are anywhere near ready for real life on a road... because the systems are deeply flawed and based on masive sets of assumptions, and occasionally that kills someone...

[Jader1973]

The BBC are reporting that the Lidar maker (Velodyne) is “baffled” by what happened because the system can see perfectly well in the dark. They’re telling their customers to carry on testing.

Goes on to say that it is then up to the Uber system to use the data to make decision: “We do not know how the Uber system of decision-making works.”.

I wonder how Uber will respond to that.

[RobDickinson]

They probably sue someone.

[AW111]

Jader1973 said:

The BBC are reporting that the Lidar maker (Velodyne) is “baffled” by what happened because the system can see perfectly well in the dark. They’re telling their customers to carry on testing.

Goes on to say that it is then up to the Uber system to use the data to make decision: “We do not know how the Uber system of decision-making works.”.

I wonder how Uber will respond to that.

Apparently neither the computer nor the driver made any attempt to brake or swerve - is the driver culpable?

re the AI thing :
The definition of "AI" varies so much between different users that it's almost meaningless.

I wrote an "expert system" in the '80s that emulated the factory diagnosis system for a car, where the operator eg. connected diagnostic termimals, entered the code flashes into the program, then performed the next step as instructed by the program.
A useful tool with no "intelligence" of any kind - however many people referred to systems of that type as "AI", and still do.

[Some Gump]

akirk said:

surely something is right or wrong, potentially wrong must make it right?!
sorry, but I have worked in IT and related worlds for 30+ years, and we are way off true AI...
conceptually it is impossible for a non-sentinent being to ever have true intelligence, unless something is conscious it can not create with personal meaning, its actions and conclusions are in a vaccuum - we can look at them externally and derive meaning because we apply our own context and beliefs, but a machine is and always will be totally incapable of approximating human intelligence, so while machines will be capable of deriving phenomenally complex conclusions, beyond the comprehension of humans, and so we will label them as AI, they can never breathe / think / philosophise / create as a human - until we gain God's skills in creation, machines will just be fast number crunchers, that does not make them true AI - the sci-fi novel fantasy of machines that can a tually think will never happen - let us not confuse complexity with originality.....

So you're saying that basically all we can ever expect to make with AI is an emulation of full on autism? That's not scary at all!

[DonkeyApple]

Jader1973 said:

The BBC are reporting that the Lidar maker (Velodyne) is “baffled” by what happened because the system can see perfectly well in the dark. They’re telling their customers to carry on testing.

Goes on to say that it is then up to the Uber system to use the data to make decision: “We do not know how the Uber system of decision-making works.”.

I wonder how Uber will respond to that.

Setting out their stall early as they know how some of their customers like to respond.

[98elise]

Tuna said:

98elise said:

Cars don't need AI. It's a rules based environment. People cannot process unexpected things very well, infact when presented with a sudden unexpected event we tend to freeze or panic.

Driving a car can be broken down into a set of rules, and when something is experienced outside of the rules then stop. A computer can do this without getting tired, enraged, looking at Facebook, doing it's makeup etc. It can also have eyes in the back of its head, or eyes that can detect objects outside our range.

There's so much misunderstanding of what the AI does in a car. It's not making choices about who or what to run over, it's making sense of the sensor data and generating the set of choices about how the car can (safely) achieve its goal. AI in these cars is deciding whether the thing at the side of the road is a tree, a bicycle, a road sign or a traffic light. The rules about what you do in each case are indeed quite simple, but first you have to 'know' what the thing is.

The same goes for traffic lanes. Our inbuilt intelligence can look at a road scene and say "yes, that's where I should go to carry straight on", but AI is needed to figure out what is a white line, what is a random road marking and what is some paint spilled by a lorry. Driverless cars are packed with AI technology to interpret what it 'sees' - and then the rules are simple. It turns out distinguishing a road sign from a KFC sign, from someone wearing a t-shirt with the Ghostbusters logo is very difficult.

That doesn't sound like AI to me. It sounds like normal rules based programming. If you showed an autonomous car an entirely new road sign it would not make a reasoned guess at what it meant.

[Jonesy23]

Certain companies think the solution involves 'deep learning' and particularly in buying their GPUs.

Real systems that actually worked shouldn't involve that sort of thing and would be much more fixed rules/decision tree based.

Basically there are a few areas of tech hype kicking around and people never fail to leap on as many bandwagons as possible especially in one part of the world. Deep learning is one of those solutions being chucked at every possible problem.

[akirk]

ash73 said:

You haven't understood what AI means, how a neural net works, or how the human brain works.

Oh I do
when you can build a human / brain, come back and explain how you do it...
if you can match the general intelligence of a child starting at school you would be generations ahead of anything out there...
If you really believe AI exists, then build it, teach it nothing and give it no framework, no rules, no logic, then put it out into the world and see what it does
There is a lot of misconception around hardware that can handle complex calculations, and millions of simultaneous processes being AI, it is not.

[AreOut]

akirk said:

Oh I do
when you can build a human / brain, come back and explain how you do it...

there are several videos on pornhub..

[Tuna]

It seems everyone on here is an expert... (not the Pornhub thing... though maybe that too)

I can't imagine how Uber could have got it so wrong when people sitting in armchairs can tell them how to do it.

[Mr2Mike]

Tuna said:

It seems everyone on here is an expert... (not the Pornhub thing... though maybe that too)

I can't imagine how Uber could have got it so wrong when people sitting in armchairs can tell them how to do it.

Who has been telling Uber how to do it? Most people are suggesting the claims made for the current technology don't stack up in the real world.

[akirk]

Mr2Mike said:

Tuna said:

It seems everyone on here is an expert... (not the Pornhub thing... though maybe that too)

I can't imagine how Uber could have got it so wrong when people sitting in armchairs can tell them how to do it.

Who has been telling Uber how to do it? Most people are suggesting the claims made for the current technology don't stack up in the real world.

exactly not possible to tell anyone how to do something which is probably several decades off being fully possible... all anyone is doing is exploding the emperors new clothes myth... I think that a lot of people are fed up of being lied to by governments and big corporates, and the PR machine for autonomous cars is a long way from reality. This sad accident merely highlights some of the discrepancies between what the manufacturers would have you believe and reality...

[Mr2Mike]

A very interesting blog from an adviser to the Google robotics team.

He suggests that the LIDAR would have seen the cyclist, if it had been working, but there is a rumour it had been disabled to test camera+radar driving (LIDAR is bulky and expensive, no one wants it if they don't absolutely need it).

And an article from the NY Times suggesting Uber have been struggling with reliability in Arizona.

Edited by Mr2Mike on Saturday 24th March 21:58

[RobDickinson]

Tesla dont equip their cars with lidar.

Just cameras and radar.

I think its possible we can do self driving without lidar, after all we do it with just stereo vision, but it is going to be harder.

Even then the radar should have spotted them, the standard volvo system probably would have.

[Mr2Mike]

RobDickinson said:

Tesla dont equip their cars with lidar.

Just cameras and radar.

And we have already seen how well that worked out...

Edited by Mr2Mike on Sunday 25th March 10:44

[RobDickinson]

Yeah nah