AV 2.0 - Another Approach to Self Driving Cars
Tech Review is running a piece on a new/recent approach to self driving, https://www.technologyreview.com/2022/05/27/1052826/ai-reinforcement-learning-self-driving-cars-autonomous-vehicles-wayve-waabi-cruise/
Four years ago, Alex Kendall sat in a car on a small road in the British countryside and took his hands off the wheel. The car, equipped with a few cheap cameras and a massive neural network, veered to the side. When it did, Kendall grabbed the wheel for a few seconds to correct it. The car veered again; Kendall corrected it. It took less than 20 minutes for the car to learn to stay on the road by itself, he says.
This was the first time that reinforcement learning-an AI technique that trains a neural network to perform a task via trial and error-had been used to teach a car to drive from scratch on a real road. It was a small step in a new direction-one that a new generation of startups believes just might be the breakthrough that makes driverless cars an everyday reality.
Reinforcement learning has had enormous success producing computer programs that can play video games and Go with superhuman skill; it has even been used to control a nuclear fusion reactor. But driving was thought to be too complicated. "We were laughed at," says Kendall, founder and CEO of the UK-based driverless-car firm Wayve.
Wayve now trains its cars in rush-hour London. Last year, it showed that it could take a car trained on London streets and have it drive in five different cities-Cambridge (UK), Coventry, Leeds, Liverpool, and Manchester-without additional training. That's something that industry leaders like Cruise and Waymo have struggled to do. This month Wayve announced it is teaming up with Microsoft to train its neural network on Azure, the tech giant's cloud-based supercomputer.
Some of the other players in this field are training their neural networks (NN) in driving simulators (still with humans as the "instructor") instead of on the road as described above.
My question is can the neural net ever get better than the person(s) that trained it? If the human (trainer) nearly misses an accident, is that what the NN will also do? Worse, I hope that they have a way of rewinding the training to some time before there is an actual accident, wouldn't want this in the training set!
I don't see that this "2.0" approach has any possibility of realizing the early hype of "zero accidents" that robot driving advocates are always going on about, but happy to hear otherwise. At best it seems like it might become nearly as good as the humans doing the training--but this would take a lot of time on the road.
Read more of this story at SoylentNews.