Amit Chakraborty - Physics Aware AI
HUMAN & AI
00:00
The Risks of Large Models in Machine Learning
For some problems, you actually need these very large models, right? Now, none of the models are. If you have verylarge models, experimentability becomes more and more complicated,. Very farability, not so much, but experimentability for sure. For example, in control, there is the concept of the eponym stability. So that takes care of this.
Transcript
Play full episode
