Re:Andrej Karpathy talks about how Tesla's NNs are structure...
by adamnemecek on Saturday November 09, 2019 @11:01PM
(#21496317)
The trick for level 5 is learning the mapping between the lidar point cloud and the video stream. It’s the best of both worlds.
[ Parent
| on HN ]
Re:Andrej Karpathy talks about how Tesla's NNs are structure...
by ben_w on Sunday November 10, 2019 @06:37AM
(#21497575)
That might be a trick. It’s not the trick human brains use. It might be equivalent to the way that when we say we want to “look at” a thing, we often also want to touch it.
https://kitsunesoftware.wordpress.com/2017/07/27/why-do-peop...
Re:Andrej Karpathy talks about how Tesla's NNs are structure...
by pgodzin on Saturday November 09, 2019 @11:04PM
(#21496325)
Tesla doesn't have have a lidar point cloud at all
Re:Andrej Karpathy talks about how Tesla's NNs are structure...
by adamnemecek on Saturday November 09, 2019 @11:04PM
(#21496327)
That’s the point. You train with lidar, deploy with cameras.
Re:Andrej Karpathy talks about how Tesla's NNs are structure...
by ojn on Sunday November 10, 2019 @02:41AM
(#21496933)
That falls apart as soon as the map and the real world deviates and you need to drive based on what’s in front of you.
Lidar helps you spot obstructions, but won’t tell you what they are and won’t help you figure out what to do to avoid them.
Want an example? Cruise’s first real world demo got stuck behind a simple taco truck in downtown SF.
Re:Andrej Karpathy talks about how Tesla's NNs are structure...
by antpls on Sunday November 10, 2019 @05:40AM
(#21497430)
The parent meant : gather training data with Lidar and Camera, then build a model with that data to learn to reconstruct a 3D space only from Camera data, and then embed that model in the cars.
Tesla is already using a model to rebuild a 3D space from Camera data only, the parent suggests to improve the quality of the transformation with high quality 3D representations from Lidar.
It's deep learning all the way down.
Re:Andrej Karpathy talks about how Tesla's NNs are structure...
by Geee on Sunday November 10, 2019 @09:51AM
(#21498125)
2D to 3D transform is simple trigonometry (using stereo / motion) and should be possible to learn without lidar. I think this is already a solved problem. One option though is to add lidars in random Teslas (e.g. 1/1000) to help with the labeling / learning.
Re:Andrej Karpathy talks about how Tesla's NNs are structure...
by adamnemecek on Sunday November 10, 2019 @01:00PM
(#21499195)
The field is called photogrammetry. It's not...precise. Point clouds have a much richer structure.
Re:Andrej Karpathy talks about how Tesla's NNs are structure...
by sheeshkebab on Sunday November 10, 2019 @08:59AM
(#21497956)
One could also train a car driving model driving in grand theft auto... but are all these tricks really what level 5 is about? I doubt