Sorry to go off-topic, but the way that font combined 'st' (with a loop back from the 't' to the top-center of the 's') was very visually distracting to me. Did that bother anyone else?
Suddenly, a leopard print sofa appears
111–114 of 114 posts
Re: Suddenly, a leopard print sofa appears
#112We spend so much effort trying to engineer intelligence, when we would get a lot farther reverse engineering intelligence. Whenever AI makes a big advance the analog was already known my neuroscientists. There is also clearly no comprehension of the importance of the topological (circuitry) defining a neural network. We always assume a fully connected network, and draw the out as such, but we don't stop to consider t…
We spend so much effort trying to engineer intelligence, when we would get a lot farther reverse engineering intelligence. Whenever AI makes a big advance the analog was already known my neuroscientists.
The problem with attempting to understand intelligence by reverse engineering the human brain is that we cannot know a priori which aspects of the human brain are necessary for intelligence to arise, and which are merely consequences/side effects of biology and chemistry. Once we discover some technique that works in a practical setting (e.g. on ImageNet), then it is fairly straightforward to find the biological analogy in the brain.
In fact, Geoff Hinton explicitly advocates an approach of "try things, keep what works, and figure out how it relates to the brain". The inverse is like finding a needle in a haystack.
There is also clearly no comprehension of the importance of the topological (circuitry) defining a neural network. We always assume a fully connected network, and draw the out as such, but we don't stop to consider that many of those Wijk interactions are completely spurious, meaning they have no information bearing role.
The purpose of training a deep neural network from data is to automatically discover what the topological circuitry of the network should be, rather than engineering it by hand. In the brain, some prior knowledge is encoded via genetics, while the rest is learned. The effect of sparsity of the weights in deep neural networks is an active area of research [1].
If you strip them away you'll start to reveal the underlying circuit at work. I've published theoretical results using artificial gene networks, but the results should be similar for ANNs.
Very interesting. If I understand correctly, the cost you are attempting to minimize is phenotypic variation, which you measure as the gross cost of perturbation (GCP). Would this cost be analogous to sensitivity to adversarial examples in the case of convolutional neural networks [2]?
Re: Suddenly, a leopard print sofa appears
#113Isn't the solution to simply have two NNs? One trained to identify leopards, another trained to identify sofas.
Regardless of how computationally expensive NNs may be now, wait a few years, and then train millions of them on different classes of objects and run them concurrently to identify new pictures.
Re: Suddenly, a leopard print sofa appears
#114OP here, and thank you kind sirs and ladies for you feedback. I'd just like to answer the recurring objection: yes, our visual experience contains a lot of frames and that seemingly refutes my MNIST example; however, you do forget about the other part of a supervised dataset, namely labels. Do we have a label provided to each thing we see in our life? Obviously not. How much time do you need to familiarize yourself w…
> How much time do you need to familiarize yourself with a new entity, like an unknown glyph or symbol?
I think you look at your subjects wrong - don't pretend your computer is an adult (which had learned most of its life) - rather consider him an infant learning letters/digits/objects for the first time. Doing so you might come across a similar learning curve to the one you have described. With the additional case we (or at least I) don't know how to bring the computer to the level of a fully grown man.
As for the problem presented in CNN, if the problem is not having the structure, why not gray-scale the structure as a secondary level for the CNN?
I'm not really from the field so excuse me if this was complete BS