MobileNet

I have a current system, right? I've...

If, if I go back to even the old days of very simple principle component analysis or support vector machines or ResNets, DenseNets, MobileNets, how did the next thing come about? You know, how did Google Net, uh, was evolved? You know, if we step in to understand what is current here and look at the limitations, then we have to imagine how do we make this better? That's the process of thinking, is understanding what we have in front of us, and then not only understanding what we have in front, but understanding how it works, what are the limitations of this current neural network, whether convolution or transformer-based, and how do we overcome the limitations? And that requires imagination.

That's requires thinking.

How do we make it better? Once we understand something, we understand the limitations, then we know what we need to do to get it to a better accuracy.