151–156 of 156 posts
Earlier quoted context omitted.
> Maybe I'm reading too much between the lines, but so far there haven't been any major successes in un-supervised learning models Word2vec, a variant of which is Transform, Google's new machine translation framework which beats the pants off every other approach, to name one. The system it is replacing is also unsupervised.
Transformer is heavily reliant on supervised learning. Not sure where you're getting that from.
Earlier quoted context omitted.
Not really, they develop it by sucking on their thumbs, drinking amniotic fluid, etc. You can see it develop, and you can see how premature babies have issues with it. Adding "breathe" to sucking and swallowing is even harder for preemies. Starting to suck when they feel something on their lips _is_ a reflex though, it isn't learnt.
reflex = pre-trained?
Back in the nineteen-nineties I'd built a neural-net creator using a method other than back-propogation that was able to create nets with about a hundred neurons that played tic-tac-toe well or perfectly (depending on the net created that day.) This was on a 12 mhz 286 using what I called static point multiplication (based on shift instructions, not multiplication instructions.) So I went to Toronto to see Professor…
My method was evolutionary, put too simply - something not completely undiscussed even at the time and it's undoubtedly happening in one way or another now. I also designed a method for parallel processing dedicated to neural nets that seemed much more efficient than just "more processors", I've no idea whether NVIDIA et al have caught up with design that or passed it by now.
I agree with the article. Throwing more cycles at a problem using the existing models isn't that answer. Better, more efficient models are needed. And yes, I'm working toward that goal.
Can you share something about your work? Or how I may get involved?
The hardest part is finding interesting challenges for the AI to conquer.
Earlier quoted context omitted.
Just ML is never going to lead to an AGI. There are lots of other things that need to be first understood about intelligence, common sense reasoning, cognition and other things which we don't have a clue of (though some think they know it all), to build an AGI.
Just ML is never going to lead to an AGI. I think there's a chance it could in principle. Or at least, I think it could yield something close. If human intelligence is largely rooted in a neural network with billions of neurons and trillions of synapses, then it seems reasonable to think that, in principle, a sufficiently deep/wide ANN could show "intelligence". But here's the rub. Doing that would be silly. We alrea…
At first glance, the obvious solution seems to be to create intelligence the same way nature did: some sort of evolution. Some sort of algorithm where multiple "networks" mutate and reproduce in some way, in response to some fitness function. Mutations that make a network more fit result in an increased reproductive rate, while other mutations decrease that rate. Evolutionary algorithms have been around for a while b…
What makes you think that our brains or intelligence is the most efficient or most effective? We happened to evolved a brain that allowed us to think on what we consider sophisticated. This doesn't mean that it's a great system. We also evolved to have a spine, which is a terrible design. There's a ton of other examples of evolution designing things that are terrible and ineffective, but get the job done. Is that wha…
My point is that perhaps we understand evolution better than we understand how to create a mind from scratch.