I was just quoting the article really.
> Neural networks are unsuccessful because they are not detecting anything. They're just making statistical correlations that happen to be right for the specific set of training data they have received.
Neural networks are moderately good at replicating the part of the brain that identifies visuals, because that is just statistics and neural networks are a slightly awkward way of doing statistics.
But only up to a point, because humans don't sit around all day looking at static images. We humans move around, and the objects we are classifying move. Our world is 4-dimensional, not 2-dimensional.
Perhaps a neural network can tell what a cat looks like, but can it tell the difference between a dead cat and a sleeping one, by watching it breathe?
Neural networks are not good at replicating the parts of the brain that deal with physical properties of the world, because that is not a statistical operation. It's an algorithmic one.
Unfortunately for them, such physical properties are not an afterthought but an inseparable part of the task they are attempting to achieve.
Yes, yes it can: http://arxiv.org/abs/1411.4389 (Long-term Recurrent Convolutional Networks for Visual Recognition and Description) It's certainly surprising how much of perception can be reduced to classification.
As someone working on neural networks, although not really on computer vision, I do not believe several of the things you said are true.
> > Pornographic images, for instance, are frequently identified not by the presence of particular body parts or structural features, but by the dominance of certain colors in the images
> Neural networks
This is definitely not what state of the art neural networks for computer vision do, although I'm sure there are many models out there that do function in that manner.
We know that neural networks learn to detect higher level features. For example, see the excellent visualizations in Matthew Zeiler and Rob Fergus' Visualizing and Understanding Convolutional Networks paper. (http://arxiv.org/pdf/1311.2901.pdf)
> Neural networks are unsuccessful because they are not detecting anything. They're just making statistical correlations that happen to be right for the specific set of training data they have received.
Are you saying that the correlations they learn aren't true on non-training data?
If so, that is false. Any competent person in ML tests on different data than they train. Well trained neural networks generalize to new data very well. That's why we use them.
> There's no robustness whatsoever. Artificial gullibility.
Are you referring to the adversarial counterexamples section of Szegedy et al's Intriguing properties of neural networks paper? (http://arxiv.org/abs/1312.6199)
This paper shows that one can craft noise which will fool a neural network. But it isn't noise that is found in real life. It's created by an optimization process which is basically looking for ways to fool the network. Basically, it discovers an optical illusion for the network.
(From the manifold perspective, it moves us in a direction perpendicular to the manifold.)