00:00 Speaker A
Now, one of the most important things we do with physical AI is first create the data to train the AI. Where does this data come from?
00:09 Speaker A
Rather than having language because we create a bunch of basic facts that we think an AI can learn from, how do we teach an AI the basic facts of physics?
00:20 Speaker A
There are lots and lots of videos out there, lots and lots of videos, but not enough to capture the variety and type of interaction we need.
00:30 Speaker A
So this is where great minds come together and turn past calculations into data.
00:43 Speaker A
Now, using synthetic data generated based on and conditioned on the laws of physics, based on and conditioned on ground truth,
00:54 Speaker A
We can now selectively and intelligently generate data and then use that data to train artificial intelligence. For example, the contents of the artificial intelligence and the Cosmos artificial intelligence world model on the left are
01:06 Speaker A
Traffic simulator output.
01:10 Speaker A
Now, this traffic simulator is not good enough for the AI ​​to learn.
01:17 Speaker A
We can put this into the Cosmos base model and generate a physically based and physically sound surround video that the AI ​​can now learn from.
