Convolutional neural network

So with convolutional neural networks, so things that powered like AlexNets, which, you know, really launched the deep learning revolution in 2012, and then ResNets and, you know, all of the, I would say the, the advancements during the 2010s was in the realm of machine learning models that were fundamentally compute bound.

You could just throw more and more FLOPS at CNNs and get better results, and you didn't really need all that much, be it memory capacity or memory bandwidth.

But it was really with the transformer, and even though the attention is all you need paper came out in 2017, I would say it did not really get the attention, you know, pun intended, uh-

Sure

But I can tell you just from my own experience, there's too much pressure from the top at traditional media outlets, you know, like

the CNNs, the Foxes, the MSNows to get the

story.

But they also have... their political agendas.

So I advised you to do matrix completion for s-- for, uh, special interpolation.

So Pui was the one getting me into CNNs, while I was still trying to smuggle classical signal processing into the project.

But that experience helped convince me that there was a very productive middle ground.

So since then, I've become particularly interested in model-based deep learning, deep unfolding, where you take an optimization algorithm and turns it in-- and, and you turn its iteration into trainable network layers, and more recently, diffusion, uh, models which, uh, which, uh, you have just described previously.

Like if we transition to dermatology, it strands a field heavily reliant on visual pattern recognition.

It is the ultimate visual medical discipline, and that makes it a prime target for automation using convolutional neural networks or CNNs.

For those of us who aren't software engineers, how exactly does a CNN look at a picture of our skin?

Well, a convolutional neural network doesn't see a mole the way a human doctor does.

It analyzes the image pixel by pixel, mathematically scanning for patterns in color, edge detection, and border irregularity.

Okay.

And in highly controlled experimental settings, you know, when fed high quality dermoscopic images which are intensely magnified and perfectly lit, these CNNs perform remarkably well.

Yeah, the numbers in the clinical reviews are definitely impressive at first glance.

And they blew it up.

And they put the official story out on both CNNs.

Back then they had Headline News and CNN and then both BBC News channels.

It was like BBC World and the other.

And they blew it up.

And they put the official story out on both CNNs.

Back then they had Headline News and CNN and then both BBC News channels.

It was like BBC World and the other.

For those listening, let's break down the two halves of this AI.

CNN stands for Convolutional Neural Networks.

And these are traditionally used in image processing.

because they are incredible at extracting spatial features.

way, the reduction in, just to mention another point, the reduction in vaccination rates occurred during the Biden administration, not now.

Okay, Bob, those last two things you said, has the local news, has any of the ABCs, CBS, CNNs brought any of that out? No.

So that's the problem.

You're talking about how many people that listen to Fox News, your station, that's who knows that shit.