Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Artificial intelligence pioneer says we need to start over

axios.com

Post reply on HN
Print view View original

21–30 of 156 posts

Re: Artificial intelligence pioneer says we need to start over

#21

Post by kobeya »

bra-ket wrote: post #13
tw1010 wrote: post #8

Earlier quoted context omitted.

Biological plausibility is just silly. What it mostly does is just put artificial constraints on the problem that doesn't need to be there. It's all mathematics. However you want to interpret or metaphorize the mathematics is up to each math-phobic field, but constraining yourself to what randomly evolved serves little good.

If the objective is to simulate general intelligence then you want to copy the natural one as closely as possible. Simulating is easier than inventing.

False dichotomy. The alternative is not blind trial and error.

Re: Artificial intelligence pioneer says we need to start over

#22

Post by iandanforth »

Brief description of the issue for non-experts:

Supervised learning

You can judge the output of your network against ground truth. You say that's a cat? Nope, it's a dog! And then slightly adjust your network so it's less likely to give that wrong answer in the future. How exactly you adjust the network is what backpropagation describes (in combination with something called a learning rate).

Unsupervised learning

You need to learn without someone telling you what the answer is. Most of the time for biological intelligence there isn't an oracle describing the truth at every moment of life to judge actions/decisions against. If you don't have someone telling you you've made a mistake, how can you know when to adjust your network? And if you don't know what the truth is, exactly how to adjust the network becomes tricky.

Somehow biological brains work without that oracle, and there are lots of ideas about how it does that. But right now for artificial intelligence none of those ideas has been shown to work so amazingly well that it has taken off like backprop has in the supervised learning world.

Hinton wants to find that amazing algorithm.

Re: Artificial intelligence pioneer says we need to start over

#23

Post by burkaman »

tw1010 wrote: post #8
bra-ket wrote: post #5

If you are interested in biologically plausible models of cognition check out 'vector symbolic architectures' and 'associative memory' research of the 90s.

Biological plausibility is just silly. What it mostly does is just put artificial constraints on the problem that doesn't need to be there. It's all mathematics. However you want to interpret or metaphorize the mathematics is up to each math-phobic field, but constraining yourself to what randomly evolved serves little good.

It's not silly to try and mimic the only known model that has ever worked. Lots of modern technology is inspired by "randomly evolved" mechanics, it's a perfectly reasonable place to start.

Re: Artificial intelligence pioneer says we need to start over

#25

Post by aaronblohowiak »

iandanforth wrote: post #22

Brief description of the issue for non-experts: Supervised learning You can judge the output of your network against ground truth. You say that's a cat? Nope, it's a dog! And then slightly adjust your network so it's less likely to give that wrong answer in the future. How exactly you adjust the network is what backpropagation describes (in combination with something called a learning rate). Unsupervised learning You…

I wager it has something to do with prediction and rules

Re: Artificial intelligence pioneer says we need to start over

#26

Post by sillysaurus3 »

iandanforth wrote: post #22

Brief description of the issue for non-experts: Supervised learning You can judge the output of your network against ground truth. You say that's a cat? Nope, it's a dog! And then slightly adjust your network so it's less likely to give that wrong answer in the future. How exactly you adjust the network is what backpropagation describes (in combination with something called a learning rate). Unsupervised learning You…

You need to learn without someone telling you what the answer is. Most of the time for biological intelligence there isn't an oracle describing the truth at every moment of life to judge actions/decisions against.

There is: Death, without reproducing. If an organism dies quickly but left children, they were biologically successful. If you didn't, then your evolutionary branch goes away.

I'm not sure this can be simulated, but it might be worth trying. Biological brains are the product of a billion years of the above strategy, so hopefully it's feasible to do the same thing.

Don't people already do that? Well, no. The strategy implies some sort of physics simulation with general rules, not a specific ruleset tuned to your domain.

Re: Artificial intelligence pioneer says we need to start over

#27

Post by twobyfour »

S_A_P wrote: post #12

So it says "In 1986, Geoffrey Hinton co-authored a paper that, four decades later, is central to the explosion of artificial intelligence." Is it just me that considers that 3 decades? I am curious which is the typo. Was it 76? or 3 decades?

1986 was 31 years ago. Maybe not decades in the sense of "1990s" being a decade. But certainly 3 periods of 10 years.

Re: Artificial intelligence pioneer says we need to start over

#29

Post by red75prime »

dnautics wrote: post #19
Eridrus wrote: post #4

Obviously a clickbait headline, few people think current tech will lead to AGI, but while we all wait on a new breakthrough in AI, we can build much more impressive systems with what we have now than even a few years ago. I went to a talk from people who had built more scalable training algorithms for Restricted Boltzmann Mahcines (which Hinton was a pioneer in) a few months ago and asked them why they had chosen to…

I don't think that's correct. There are a lot of people who should know better that think that the current ml technology will lead to agi.

Let me guess. You think so, because we don't know what the general intelligence is, but the current approaches of ML are surely can't be extended to produce it.

Re: Artificial intelligence pioneer says we need to start over

#30

Post by striking »

sillysaurus3 wrote: post #26
iandanforth wrote: post #22

Brief description of the issue for non-experts: Supervised learning You can judge the output of your network against ground truth. You say that's a cat? Nope, it's a dog! And then slightly adjust your network so it's less likely to give that wrong answer in the future. How exactly you adjust the network is what backpropagation describes (in combination with something called a learning rate). Unsupervised learning You…

You need to learn without someone telling you what the answer is. Most of the time for biological intelligence there isn't an oracle describing the truth at every moment of life to judge actions/decisions against. There is: Death, without reproducing. If an organism dies quickly but left children, they were biologically successful. If you didn't, then your evolutionary branch goes away. I'm not sure this can be simul…

>> Most of the time for biological intelligence there isn't an oracle describing the truth at every moment of life to judge actions/decisions against. [emphasis mine]

> There is: Death, without reproducing.

The problem is that it's not describing the truth at every moment of life, just the total approximate fitness of that lifeform by the end of its lifetime. It would be time consuming and cost-inefficient to simulate many possible configurations of actors until the end of their lifetimes.

Post reply on HN