Nvidia on new self-driving system: “basically 5 years ahead and coming in 2017” • Hacker News Board

Re: Nvidia on new self-driving system: “basically 5 years ahead and coming in 2017”

#111

Post by sliken » Sun, Nov 13, 2016, 1:47 AM UTC

kchoudhu wrote: post #25

bluthru wrote: post #22

Earlier quoted context omitted.

What's a hard case for vision? That's how our eyes work. LiDAR has problems with precipitation too.

Depth map extraction from vision in real time depends on accurate algorithmic merging of past frames, color gradients and motion vector extraction to come up with a 3D map of what's around the vehicle. Contrast this with LIDAR, which can present a depth map in real time by sending out an array of light pulses and timing how long they take to come back to the IP. Which method is more likely to have implementation erro…

In a perfect world lidar is far superior. It's trivial to say "out of the 50M voxels returned, are any from where the car will be in 3 seconds". So lidar wins in the best case, maybe even the average case.

However in the real world where the car/sensors gets dirty/wet, and the air is filled with snowflakes, raindrops, or mist it gets much more complicated.

Even if statistically better there's value to acting more like a human driver. After all the roads are filled with human controlled cars and any deviation from human norms because of weather(like wind, snow, rain, fog, and blown sand) cause problems.

Re: Nvidia on new self-driving system: “basically 5 years ahead and coming in 2017”

#112

Post by sliken » Sun, Nov 13, 2016, 1:54 AM UTC

varelse wrote: post #36

maxerickson wrote: post #24

Earlier quoted context omitted.

Is 10 watts the correct figure? The powertrain uses 300+ Watts, so it seems like a manageable number if it is correct.

Seems indeed, and in fact an A/C will eat a kilowatt or more, but... The components of this system need to generate next to zero heat and they need to be placed in all sorts of inconvenient locations. That's causing automakers to desire extremely low-power dedicated circuitry over GPUs. Consider as an example the C7 Corvette: despite its enormous blindspot, it still doesn't have blindspot indicators because they can'…

I don't believe that. Today's cameras are tiny and the C7 isn't a particularly small car.

Re: Nvidia on new self-driving system: “basically 5 years ahead and coming in 2017”

#113

Post by kpil » Sun, Nov 13, 2016, 1:54 AM UTC

viraptor wrote: post #108

Cybiote wrote: post #85

Earlier quoted context omitted.

> People drive reasonably well using vision primarily This is not accurate however. Other important senses in use include proprioceptive, hearing and tactile feedback from wheels. In addition to vision and the improved dynamic range of eyes, there is the important fact that human vision integrates a world model into expectations. Human vision also models time and motion which help manage where to focus attention. Hum…

> This is true but only in a limited sense. Isn't what we need very limited anyway? What the cars need is recognition of obstacles and the type of the obstacle in a very limited range. Basically when something is on the road, it doesn't matter whether it's a moose or a deer - you slow down and avoid, or brake depending on the environment. "what is in this picture" classifiers don't seem like a good algorithm to use i…

Eh. Is it clouds or a truck, or a shadow on the road or a moose? A tree besides the road or a cyclist) Hard to tell without really good classifiers.

I guess that with a lidar or some kind of 3d perception, you can relax the demands on the classifier a bit and ignore really flat things and the sky...

Re: Nvidia on new self-driving system: “basically 5 years ahead and coming in 2017”

#114

Post by Cybiote » Sun, Nov 13, 2016, 1:55 AM UTC

enraged_camel wrote: post #102

Cybiote wrote: post #85

Earlier quoted context omitted.

> People drive reasonably well using vision primarily This is not accurate however. Other important senses in use include proprioceptive, hearing and tactile feedback from wheels. In addition to vision and the improved dynamic range of eyes, there is the important fact that human vision integrates a world model into expectations. Human vision also models time and motion which help manage where to focus attention. Hum…

>>This is not accurate however. Other important senses in use include proprioceptive, hearing and tactile feedback from wheels. In addition to vision and the improved dynamic range of eyes, there is the important fact that human vision integrates a world model into expectations. Human vision also models time and motion which help manage where to focus attention. Humans can additionally predict other agents and other…

Very much so and that's why we don't want our cars to perfectly replicate humans. More sensors plus a limited and narrow AI is better than just vision and a smarter AI.

Re: Nvidia on new self-driving system: “basically 5 years ahead and coming in 2017”

#115

Post by sliken » Sun, Nov 13, 2016, 1:59 AM UTC

sushirain wrote: post #92

ilaksh wrote: post #41

Earlier quoted context omitted.

What fundamental problems remain to be solved for self driving cars?

Vehicle detection: detect any vehicle, even from the side, even if its shape is rare, even at dark, etc. This could be solved with data. Control: when to yield, without watching the face of the other driver, etc. Edge cases: obeying a police officer, yielding to an ambulance, cooperating with other cars.

Nvidia mentioned identifying humans, dogs, parked cars, trucks, street signs, street lights, etc better than a human.

On their todo list was identifying kinda of cars (like a police car), kinds of trucks (school bus and ambulance), and acting appropriately.

I've seen google mention responding to hand gestures from bicyclist. But also being exceedingly polite and repeatedly stopping as a track standing bicyclist rides slightly backwards and forwards.

Re: Nvidia on new self-driving system: “basically 5 years ahead and coming in 2017”

#116

Post by mikeash » Sun, Nov 13, 2016, 2:12 AM UTC

closeparen wrote: post #91

guiambros wrote: post #86

Earlier quoted context omitted.

> ... humans don't need LIDAR to drive, why should computers? That being the case, wouldn't we be limiting self-driving technology to the same traffic-related death rates as humans? Maybe 10, 20% better, but still fundamentally close. For self-driving cars to be truly successful, the death rates will need to be an order of magnitude better. An incremental improvement won't convince governments and the public at large…

>Maybe 10, 20% better, but still fundamentally close. You are asserting that human drivers are essentially perfect, because in 80-90% of their crashes, the information necessary to avoid the collision just isn't available visually. That seems like an incredibly optimistic view of human drivers. Collisions happen because a driver does not look at, see, understand, or act appropriately on available visual signals. Or t…

A huge number of traffic deaths are due to alcohol. An autonomous system that's as safe as a sober human would improve safety by a factor of 2 or 3. Many of the other deaths are due to distraction, inattention, or slow reaction times. Get rid of those and you can probably see an order of magnitude improvement with something that is nominally "no better than a human driver."

Re: Nvidia on new self-driving system: “basically 5 years ahead and coming in 2017”

#117

Post by Hydraulix989 » Sun, Nov 13, 2016, 2:29 AM UTC

dave_sullivan wrote: post #109

Hydraulix989 wrote: post #95

Earlier quoted context omitted.

Well, not with traditional feedforward networks (LeNet, etc.). You can't run the classifier and find tires, then wheels, and then a car; but you do get composition of features.

> not with traditional feedforward networks (LeNet, etc.) I'd argue they are implicitly doing this. > You can't run the classifier and find tires, then wheels, and then a car; Why can't you run a classifier for tires, one for wheels, one for cars, then combine their outputs for a final classifier maybe based on a decision tree? You can train all the networks at the same time and it will give you a probability distrib…

Doesn't scale. You don't have N brains to compose every representation.

Re: Nvidia on new self-driving system: “basically 5 years ahead and coming in 2017”

#118

Post by Cybiote » Sun, Nov 13, 2016, 2:48 AM UTC

dave_sullivan wrote: post #94

Cybiote wrote: post #85

Earlier quoted context omitted.

> People drive reasonably well using vision primarily This is not accurate however. Other important senses in use include proprioceptive, hearing and tactile feedback from wheels. In addition to vision and the improved dynamic range of eyes, there is the important fact that human vision integrates a world model into expectations. Human vision also models time and motion which help manage where to focus attention. Hum…

> The key difference is learning in animals occurs by breaking things down in terms of modular concepts, so even when things are not recognized new things can be labeled as a composition of smaller nearby concepts. Machines cannot yet do this well at all and certainly not as flexibly. Actually, that's pretty much what deep learning is doing. For instance: https://papers.nips.cc/paper/5027-zero-shot-learning-through..…

I'm talking about learning modularly. Children can categorize things they've never seen before by inventing labels on the spot, they're not limited to selecting from a preexisting set (e.g. a lion is a "big cat"). They can recognize novelty and ask questions if nothing they know quites fit.

As a human, you are able to learn the general concept of leg and understand it, even if it is in a context you have never seen before and an object you've never seen or never seen used in that way before. Everything you learn, is also as part of a set of relations. Each of individual concept modified in a precise manner as you learn something new about any of them. A big part of human intelligence, from the simple naming of things, to the highest levels of science is taking parts of things you know and putting them together in novel ways.

In Neural nets this: https://arxiv.org/abs/1511.02799 is in line of what I mean.

Re: Nvidia on new self-driving system: “basically 5 years ahead and coming in 2017”

#119

Post by pistle » Sun, Nov 13, 2016, 3:18 AM UTC

First, I'm all for AI-based transportation solutions, but why does it seem like there aren't nearly enough redundancies being considered? There are just going to be competing proprietary solutions?

In the US, the NHSTA should get ahead of things and push for open standards and potentially for some level of development of standard safety features like being able to set some form of material or device on objects to mark them in ways that can transmit specific information about objects like other cars, fixed structures, etc.

I could make 10's of millions selling stickers or paint additives that would mark a human-driven car's edges to help "protect" it from automated vehicles' AI.

Re: Nvidia on new self-driving system: “basically 5 years ahead and coming in 2017”

#120

Post by varelse » Sun, Nov 13, 2016, 3:47 AM UTC

dogma1138 wrote: post #53

varelse wrote: post #21

NVIDIA's Drive PX 2 has too high power consumption and too low perf/W for the moment. They're winning this space because they are there more than that they are the best possible solution. And they may continue to win because successful execution of an 80% product is worth far more than a 90+% powerpoint processor cough TenSilica et al. cough , or because this is such a huge potential market, it might actually go to a…

>I'm not a big fan of $150K high-end servers filled with $5000 GPUs that can be bested with clever code on a $25K server fill with $1200 consumer GPUs. But I am a huge fan of charging what you can while you are unopposed. It's just that I think that state is temporary. There is virtually not a single "enterprise" grade product which can't be made at least 50% cheaper (or sometimes 10 times...) with off the shelf cons…

Virtualization? Don't care, in fact, virtualization is what disabled P2P copies and created craptastic upload/download perf on AWS until the P2 instance.

DMA/CUDA Direct? Say Hello to P2P and staged MPI transfers, faster, cheaper (and usually better). Know your PCIE tree FTW.

Support? As someone who has been playing with GPUs for over a decade, bugs get fixed in the next CUDA release no matter Tesla or GeForce, if ever.

$100M service? Yep I'm with you. But I prefer a world without a huge barrier to entry to building that service, especially a barrier built 99% on marketecture. I want to build on commodity hardware and deploy in the datacenter.

Unfortunately, sales types seem to hate that outlook.