Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Another Tesla on autopilot steers towards a barrier

reddit.com

Post reply on HN
Print view View original

321–330 of 481 posts

Re: Another Tesla on autopilot steers towards a barrier

#321

Post by saagarjha »

jacquesm wrote: post #161
aequitas wrote: post #160

Earlier quoted context omitted.

Don't navigation systems 'snap' to the most probable lane you are driving in based on general direction instead of your exact GPS coordinates? A good example of this I think is when driving in tunnels, there is no GPS info there, but the navigation shows you following the (curved) road.

Yes, they do that, and they rarely get that wrong. Also, the better ones have inertial backup & flux gate for when GPS is unavailable. Inertial backup could also help to detect the signatures of lane changes and turns.

Try driving near a place with a lot of roads that are close together, like an airport. You’ll quickly see that snapping doesn’t really work, especially if you deviate from the route the GPS is trying to take you on.

Re: Another Tesla on autopilot steers towards a barrier

#322

Post by gwern »

DSingularity wrote: post #182
dancek wrote: post #165

Earlier quoted context omitted.

> It's not possible to engineer an autonomous system that never fails, but it is possible to engineer one in such a way that it never fails to detect that it has failed. This is a very good point: just like a human driver should slow down if they can't observe the road ahead well enough, an AI should slow down when it's not confident enough of its surroundings. This is probably very difficult to do, and I'm skeptical…

I feel that’s the problem. Techniques like SVM can provide reasonable definitions of confidence. Reinforcement Learning.. DNN... the mechanisms behind autonomous cars... are they able to do this??

> Reinforcement Learning.. DNN... the mechanisms behind autonomous cars... are they able to do this??

Sorta. There are ways to extract kinds of uncertainty and confidence from NNs: for example, Gal's dropout trick where you train with dropout and then at runtime you use an 'ensemble' of multiple dropout-ed versions of your model, and the set of predictions gives a quasi-Bayesian posterior distribution for the predictions. NNs can be trained directly via HMC for small NNs, and there are arguments that constant-learning-rate SGD 'really' implements Bayesian inference and an ensemble of checkpoints yields an approximation of the posterior, etc. You can also train RL NNs which have an action of shortcutting computation and kicking the problem out to an oracle in exchange for a penalty, which trains them to specialize and 'know what they don't know' so they choose to call the oracle when they're insufficiently sure (this can be done for computational savings if the main NN is a small fast simple one and the oracle is a much bigger slower NN, or for safety if you imagine the oracle is a human or some fallback mechanism like halting).

I have some cites on these sorts of things in https://www.gwern.net/Tool-AI and you could also look at the relevant tags https://www.reddit.com/r/reinforcementlearning/search?q=flai... and https://www.reddit.com/r/reinforcementlearning/search?q=flai...

Re: Another Tesla on autopilot steers towards a barrier

#323

Post by xapata »

mhermher wrote: post #280
DSingularity wrote: post #182

Earlier quoted context omitted.

I feel that’s the problem. Techniques like SVM can provide reasonable definitions of confidence. Reinforcement Learning.. DNN... the mechanisms behind autonomous cars... are they able to do this??

not an expert, but aren't the output nodes thresholded to make a decision? How far you are from the threshold might be interpretable as confidence possibly?

For a single perceptron, not for the network as a whole.

Re: Another Tesla on autopilot steers towards a barrier

#324

Post by dragonwriter »

TaylorAlexander wrote: post #209
raisedbyninjas wrote: post #197

Earlier quoted context omitted.

>This is actually the most serious aspect of all of these crashes: the system does not seem to be aware when it is getting things wrong. Although serious, this is working as designed. Level II self-driving doesn't have automation that makes guarantees about recognizing scenarios it cannot handle. At level III, the driver can safely close their eyes until the car sounds the alarm that it needs emergency help. Audi pla…

I disagree. It’s not as though level II is a hard and fast definition - Tesla wants level 4 and claims the cars possess the hardware for that already. So you’d think their level II would still be smart enough to detect these problems to some degree. It’s not a fixed system but one they keep upgrading. I would expect this nominally level II system to have more smarts than a system that is designed never to exceed leve…

> Tesla wants level 4 and claims the cars possess the hardware for that already.

You can't possibly determine what hardware is required for Level 4 until you have proven a hardware/software combination, so that's just empty puffery, but even if it was true...

> So you’d think their level II would still be smart enough to detect these problems to some degree.

No, because the smartness of their system is about the software. They could have hardware sufficient to support Level 4 autonomy and better-than-human AI running software that only supports a less-thsn-adequate version of Level 2 autonomy. What their hardware could support (even if it was knowable) gives you no basis for belief about what their current software on it supports, except that it won't exceed the limits set by the hardware.

Re: Another Tesla on autopilot steers towards a barrier

#325

Post by lisper »

dorfsmay wrote: post #310
lisper wrote: post #122

Here is my armchair diagnosis: right before the car veers towards the barrier it drives through a stretch of road where the only visible lane marker is on the left. Then the right lane marker comes into view at about the point where the lane starts to widen out for the lane split. The lines that will become the right and left lane markers of the split left and right lanes respectively are right next to the van in fro…

> my armchair diagnosis > my 1991 PH.D. thesis on autonomous driving Going to hide in a corner and stay quiet on HN until I forget about this comment!

:-)

Just in case you're interested:

https://vtechworks.lib.vt.edu/handle/10919/38880

and the associated conference paper:

http://www.flownet.com/gat/papers/aaai92.pdf

Most of the work was done on a Mac II with 8MB (that's megabytes, not gigabytes) of RAM.

The progress that has been made since those days boggles my mind.

Re: Another Tesla on autopilot steers towards a barrier

#326

Post by m4x »

I don't understand why people use Autopilot, or why Tesla has released it.

To use it safely and according to Tesla's instructions, you have to remain 100% vigilant at all times - so you might as well just drive yourself.

And if you fail to remain vigilant, which is likely since you are sitting passively in the drivers seat, you might kill somebody.

Where's the upside? Why on Earth would I want to use such a product?

Re: Another Tesla on autopilot steers towards a barrier

#327

Post by kosievdmerwe »

oldgradstudent wrote: post #266
swiley wrote: post #57

Earlier quoted context omitted.

And this is why if I ever get a self driving car I want: 1) root 2) control over updates 3) everything to be completely open

Everything to be completely open is really important, but I don't want you to have root or control over update if you are using your car on a public road.

Only certified (by the government) updates should be allowed on self driving cars. It should be at least a misdemeanor to have your own code on the car.

Re: Another Tesla on autopilot steers towards a barrier

#328

Post by icc97 »

jacquesm wrote: post #109
slivym wrote: post #66

Earlier quoted context omitted.

I've got to take issue with you there. Tesla Engineering knows better than anybody what updates they did and how often they update etc. Tesla PR knows nothing about what updates the engineering team did. At least some people in Tesla PR probably don't even know the cars update their software regularly. It's bad practice for them to speak out of turn, but I can absolutely see the PR team not having a good grasp of wha…

I'm sorry but Tesla PR == Tesla. If they don't have a good grasp on this they should STFU until they do. That would make Tesla even worse in my book. Their job is not to put out the best numbers possible, their job is to inform. Most likely they were more worried about the effect of their statement on their stock price than they were worried about public safety. If they do put out numbers (such as 85K trips past that…

Well I kind of came to hope that Tesla PR was less BS than other PR companies. But that memo basically shows they're no different from all other PR, just twisting the truth to avoid negative stories until the fuss dies down.

Re: Another Tesla on autopilot steers towards a barrier

#329

Post by lisper »

aeorgnoieang wrote: post #313
lisper wrote: post #122

Here is my armchair diagnosis: right before the car veers towards the barrier it drives through a stretch of road where the only visible lane marker is on the left. Then the right lane marker comes into view at about the point where the lane starts to widen out for the lane split. The lines that will become the right and left lane markers of the split left and right lanes respectively are right next to the van in fro…

> It's not possible to engineer an autonomous system that never fails, but it is possible to engineer one in such a way that it never fails to detect that it has failed. That is an extremely surprising result. How is that possible? Are you really claiming that any control system can be engineered to detect that is has failed in any possible manner? What's an example of actual real-world system like that?

> Are you really claiming that any control system can be engineered to detect that is has failed in any possible manner?

No, of course not. But it is possible to reduce the probability of non-cognizant failure to arbitrarily low levels -- at the expense of cost and the possibility of having a system that is too conservative to do anything useful.

Re: Another Tesla on autopilot steers towards a barrier

#330

Post by lisper »

mannykannot wrote: post #319
lisper wrote: post #292

Earlier quoted context omitted.

You've done a great job of describing the problem. It is manifestly possible to drive missed detection rates very close to zero without too many false alarms because humans are capable of driving safely.

Ah, yes... you have convinced me - though humans can apply generalized intelligence to the problem, which I imagine is particularly useful in lots of special cases.

General intelligence is not needed. Illiterate people can drive. People who can't do math can drive.
Post reply on HN