Physical AI Lab // Episode 0.4 // Quiz
Ten thousand tries in a virtual world
What simulation is, and why robots can't do without it
01
What is the most accurate way to understand "simulation" (a physics engine) as the video uses it?
A recording of a real robot
A "fake world" in a computer, calculated by the laws of physics
A new kind of robot
A dataset
02
What is the first benefit of "practising in a virtual world" in the video?
Safety — breaking things costs nothing
Cheaper electricity
Nicer graphics
No need to write code
03
Where does the claim "learns to walk in under four minutes" come from?
A maker's advert
A paper (thousands of robots training in parallel in simulation)
An offhand remark
TV news
04
How should you read "moving simulation onto graphics cards made it a hundred to a thousand times faster"?
It is always this fast
It's an officially guaranteed number
It's a technical report's result, compared against "the conventional pipeline with CPU simulation and GPU neural networks"
It's a measured average
05
According to the video, what is simulation's biggest weakness?
It calculates too slowly
It can't train many at once
It has no gravity
It isn't quite the same as reality; what's learned must be tuned again on a real robot
Check