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

Learning from context is harder than we thought

hy.tencent.com

Post reply on HN
Print view View original

71–80 of 140 posts

Re: Learning from context is harder than we thought

#71

Post by em-bee »

nimonian wrote: post #65
em-bee wrote: post #20

Earlier quoted context omitted.

and what is this quote supposed to explain? that language prevents a child from learning nuance? sounds like nonsense to me. a child first learns broad categories. for example some children as they learn to speak think every male person is dad. then they recognize everyone with a beard is dad, because dad has a beard. and only later they learn to differentiate that dad is only one particular person. same goes for the…

Wittgenstein famously said "The limits of my language mean the limits of my world." Alan Watts suggests people like Wittgenstein should occasionally try to let go of this way of thinking. Apologies if it is sentimental but I hope you'll give him a chance, it's quite short: https://m.youtube.com/watch?v=heksROdDgEk In reflection of all of this, I think that the quote you're responding to only meant to say that experie…

the way i read the quote it felt less like building an abstraction and more like destroying the richness.

but abstractions are mere shortcuts. but everything is an abstraction. to counter wittgenstein, language is not actually limited. we can describe everything to the finest detail. it's just not practical to do so every time.

physics, chemistry, we could describe a table as an amount of atoms arranged in a certain way. but then even atom is an abstraction over electrons, protons and neutrons. and those are abstractions over quarks. it's abstractions all the way down, or up.

language is abstractions. and that fits well with your meditation example. stop thinking -> remove the language -> remove the abstractions.

Re: Learning from context is harder than we thought

#72

Post by api »

nemomarx wrote: post #47
cs702 wrote: post #33

Earlier quoted context omitted.

Our brains, which are organic neural networks, are constantly updating themselves. We call this phenomenon "neuroplasticity." If we want AI models that are always learning, we'll need the equivalent of neuroplasticity for artificial neural networks. Not saying it will be easy or straightforward. There's still a lot we don't know!

How would you keep controls - safety restrictions - Ip restrictions etc with that, though? the companies selling models right now probably want to keep those fairly tight.

This is why I’m not sure most users actually want AGI. They want special purpose experts that are good at certain things with strictly controlled parameters.

Re: Learning from context is harder than we thought

#73

Post by 4b11b4 »

cs702 wrote: post #33
4b11b4 wrote: post #31

Earlier quoted context omitted.

I'm not sure if you want models perpetually updating weights. You might run into undesirable scenarios.

Our brains, which are organic neural networks, are constantly updating themselves. We call this phenomenon "neuroplasticity." If we want AI models that are always learning, we'll need the equivalent of neuroplasticity for artificial neural networks. Not saying it will be easy or straightforward. There's still a lot we don't know!

I wasn't explicit about this in my initial comment, but I don't think you can equate more forward passes to neuroplasticity. Because, for one, simply, we (humans) also /prune/. And... Similar to RL which just overwrites the policy, pushing new weights is in a similar camp. You don't have the previous state anymore. But we as humans with our neuroplasticity do know the previous states even after we've "updated our weights".

Re: Learning from context is harder than we thought

#76

Post by 4b11b4 »

api wrote: post #72
nemomarx wrote: post #47

Earlier quoted context omitted.

How would you keep controls - safety restrictions - Ip restrictions etc with that, though? the companies selling models right now probably want to keep those fairly tight.

This is why I’m not sure most users actually want AGI. They want special purpose experts that are good at certain things with strictly controlled parameters.

I agree, the fundamental problem is we wouldn't be able to understand it ("AGI"). Therefore it's useless. Either useless or you let it go unleashed and it's useful. Either way you still don't understand it/can't predict it/it's dangerous/untrustworthy. But a constrained useful thing is great, but it fundamentally has to be constrained otherwise it doesn't make sense

Re: Learning from context is harder than we thought

#77

Post by prng2021 »

cs702 wrote: post #24

The problem is even more fundamental: Today's models stop learning once they're deployed to production. There's pretraining, training, and finetuning, during which model parameters are updated. Then there's inference, during which the model is frozen. "In-context learning" doesn't update the model. We need models that keep on learning (updating their parameters) forever, online, all the time.

Thanks for repeating what the author explained.

Re: Learning from context is harder than we thought

#78

Post by api »

4b11b4 wrote: post #76
api wrote: post #72

Earlier quoted context omitted.

This is why I’m not sure most users actually want AGI. They want special purpose experts that are good at certain things with strictly controlled parameters.

I agree, the fundamental problem is we wouldn't be able to understand it ("AGI"). Therefore it's useless. Either useless or you let it go unleashed and it's useful. Either way you still don't understand it/can't predict it/it's dangerous/untrustworthy. But a constrained useful thing is great, but it fundamentally has to be constrained otherwise it doesn't make sense

The way I see it, we build technology to be what we are not and do what we can’t do or things we can do but better or faster.

An unpredictable fallible machine is useless to us because we have 7+ billion carbon based ones already.

Re: Learning from context is harder than we thought

#79

Post by Izkata »

BobbyTables2 wrote: post #69
cs702 wrote: post #24

The problem is even more fundamental: Today's models stop learning once they're deployed to production. There's pretraining, training, and finetuning, during which model parameters are updated. Then there's inference, during which the model is frozen. "In-context learning" doesn't update the model. We need models that keep on learning (updating their parameters) forever, online, all the time.

How long will it take someone to poison such a model by teaching it wrong things? Even humans fall for propaganda repeated over and over . The current non-learning model is unintentionally right up there with the “immutable system” and “infrastructure as code” philosophy.

> How long will it take someone to poison such a model by teaching it wrong things?

TayTweets was a decade ago.

Re: Learning from context is harder than we thought

#80

Post by pankajdoharey »

embedding-shape wrote: post #38
cs702 wrote: post #24

The problem is even more fundamental: Today's models stop learning once they're deployed to production. There's pretraining, training, and finetuning, during which model parameters are updated. Then there's inference, during which the model is frozen. "In-context learning" doesn't update the model. We need models that keep on learning (updating their parameters) forever, online, all the time.

> We need models that keep on learning (updating their parameters) forever, online, all the time. Do we need that? Today's models are already capable in lots of areas. Sure, they don't match up to what the uberhypers are talking up, but technology seldom does. Doesn't mean what's there already cannot be used in a better way, if they could stop jamming it into everything everywhere.

Continuous learningin current models will lead to catastrophic forgetting.
Post reply on HN