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

DeepMind says reinforcement learning is ‘enough’ to reach general AI

venturebeat.com

Post reply on HN
Print view View original

151–160 of 312 posts

Re: DeepMind says reinforcement learning is ‘enough’ to reach general AI

#151

Post by zcw100 »

rocgf wrote: post #81
for_i_in_range wrote: post #74

Earlier quoted context omitted.

“Show me the incentives I’ll show you the outcome.” Google make money, Google Bad. Deep Mind owned by Google, Deep Mind bad! The above conclusion stands trite. Perhaps the inverse is true. Google and Deep Mind, if correct, could be hurting themselves more than helping themselves. Why? Creating a future species who’s too smart to click on ads, and too smart to remain subject to its whims, doesn’t sound like it’d be go…

No offence, but I think you are extremely wrong. Creating an AGI is the endgame for everything. Who cares about ads when you have an AI that can learn to do anything and improve upon itself continuously?

Really? I created two GI's and it wasn't very hard and was actually quite fun. Training them is a bit of a pain though. I'm willing to bet that based on total calories consumed they are amazingly efficient compared to their hypothetical AGI counterparts.

Re: DeepMind says reinforcement learning is ‘enough’ to reach general AI

#152

Post by harry8 »

chrisco255 wrote: post #83
A-Train wrote: post #76

Earlier quoted context omitted.

The algorithms to train, initialize the networks, new architectures are far more important than the hardware advances. If people knew how to train NNs 50 years ago we would live in a different world.

We did. they just didn't have the same computation abilities back then.

I find it really interesting that when Richard Feynman did a sabbatical at Thinking Machines when they were developing the early parallel execution hardware that's really not worlds away from modern GPUs he got them in touch with one of the leading neural network theorists as an obvious use for the tech. When he wasn't fixing their hardware designs using systems of differential equations.

It would be an interesting thing to know more about.

Re: DeepMind says reinforcement learning is ‘enough’ to reach general AI

#153

Post by croon »

tinco wrote: post #149
croon wrote: post #146

Earlier quoted context omitted.

The interface to the car is a solved problem. > You can read a book on how to hit a ball with a baseball bat, you can even practice and get good at it, but that still doesn't mean you would actually be able to hit a ball thrown by a professional pitcher. If I hade incredibly fast reflexes and actuators I could.

Similarly, DeepMind's software might be able to drive a car, would it have a similar neuron count, connectivity, perception systems and training you received. Or maybe it couldn't, because the software is not as efficient as the organisation of your brain is. Or because there's hardcoded routines evolved in your brain that it lacks. What I'm saying is that just that because an AGI can't drive a car, it doesn't mean i…

> Similarly, DeepMind's software might be able to drive a car, would it have a similar neuron count, connectivity, perception systems and training you received.

Admittedly I'm a layman in this area, but could it? AFAICT it would only work on trained set data and whatever generalizations can be made on that and not infer unseen scenarios like humans do readily.

> What I'm saying is that just that because an AGI can't drive a car, it doesn't mean it's an AGI.

I understood what you meant from your first post, I'm simply disagreeing on account of the very definition of AGI.

You can't have an ameoba level AGI and still call it (a limited) AGI. Either it can understand/learn any human task, or it can't.

The definition is made for a reason. Watering it down for any specific generation of AI serves no benefit.

Re: DeepMind says reinforcement learning is ‘enough’ to reach general AI

#154

Post by neltnerb »

exporectomy wrote: post #129
neltnerb wrote: post #4

Earlier quoted context omitted.

Yes, it's easy to be convinced on either side, the arguments write themselves. Yes, eventually a learning system might learn enough to be indistinguishable from intelligence. Or this might be entirely the wrong path and detracting from genuine new innovations in how we think about AI. We won't be able to tell whether it's AGI or just good enough at trained tasks to trick us.

It can prove its intelligence by making testable predictions of the future better than us. As for whether it's "real" AGI or just acts like it, doesn't really matter. I think the Chinese room problem has been agreed on as not a problem, hasn't it?

My last sentence was a statement of that problem, not a question.

Re: DeepMind says reinforcement learning is ‘enough’ to reach general AI

#155

Post by harry8 »

rusk wrote: post #42
harry8 wrote: post #18

Some Bozo who has heard all this many times before is suspicious of claims from places like Deep Mind who have a financial incentive to make them (keep funding) where there aren't working machines to back that claim up. Some Bozo has no credentials, no reputation, no track record of publications and barely supports the claim they're making with anything much. Some Bozo has no financial incentives or otherwise to opin…

Edit: sorry just realised you’re making the same point as me more or less. Putting yourself in third person. I’ll let my comment stand anyhow :) Screwing my face up, looking at this sideways … but it seems as though you’re saying that the Bozos of HN have nothing useful to contribute to this discussion based on … [rereads] … their lack of academic credentials in the area… you could say this about just about any HN po…

Not quite the same point. Yep some bozo is me but needn't be. There's plenty who share that suspicion of AI research but have little else in common. And all of us may be wrong for different reasons.

The vague point was to show someone with zero reputation, credentials, specific expertise in the field or anything much seems to be pretty convincing in response to this hugely funded ivory tower exercise by spitting, cocking an eyebrow and saying "So you think so, eh? Wanna bet?"

This is a statement about the state of AI research credibility. Do you feel the first breezes of a deep AI winter coming on? (I don't know, I'm disinterested but not uninterested. Rising tides lift all ships etc. And vice versa). Neutral nets are cool. Is all ML a bit overrated? Is learning a misleading name to give to applied statistics?

I don't have answers, just suspicions. I could be very wrong, of course.

Re: DeepMind says reinforcement learning is ‘enough’ to reach general AI

#156

Post by Planktonne »

andyxor wrote: post #69
eeegnu wrote: post #29

Earlier quoted context omitted.

> towards understanding natural intelligence, a prerequisite for creating an artificial one. I've thought about this before, and I'm not convinced it's really prerequisite. Naturally developed intelligence in my mind may actually be highly constrained and inefficient because it was limited to what was biologically feasible. i.e. There may be simpler ways of achieving comparable results. Natural intelligence does howe…

yes, it's a recurring thread, "do we really need to mimic the birds in order to build airplanes", etc. I think someone serious about AI should treat it not as engineering problem but as a science, like physics, which starts with model of nature, and experiment to prove or disprove the theory. Nature provides the constraints by which theory is developed, which radically limits the "search space" of theories. Otherwise…

Mimicking birds wasn't necessary for flight.

However, understanding them absolutely was; we didn't end up taking exactly the same route to the sky, but we absolutely learnt from birds on the way.

Re: DeepMind says reinforcement learning is ‘enough’ to reach general AI

#157

Post by harry8 »

graderjs wrote: post #95
harry8 wrote: post #18

Some Bozo who has heard all this many times before is suspicious of claims from places like Deep Mind who have a financial incentive to make them (keep funding) where there aren't working machines to back that claim up. Some Bozo has no credentials, no reputation, no track record of publications and barely supports the claim they're making with anything much. Some Bozo has no financial incentives or otherwise to opin…

Plot-twist: Deep Mind comes out as Some Bozo comment author

Well my writing style has been the subject of abuse on this website in the precise form that it resembles words generated by a bad algorithm.

"Only a true AI would deny their being."

Re: DeepMind says reinforcement learning is ‘enough’ to reach general AI

#158

Post by mirker »

rich_sasha wrote: post #101
d110af5ccf wrote: post #78

Earlier quoted context omitted.

Deep Mind is cutting edge ML in general, right? Doesn't Google actively apply the lessons learned all over the place? YouTube content recommendation stands out to me in particular. Translation and automated closed captioning are also obviously ML based. I'd guess that most of the really interesting stuff would be behind the scenes and not immediately visible to end users though.

I’m stressing the business part. YouTube is a loss-making business year after year. Deep Mind gloss doesn’t seem to change that. If indeed it even is Deep Mind making those improvements, Google has lots of other ML groups, such as Google Brain, and these are more directly focused on Google products. There’s no denying their academic success, or game playing etc, but as far as I can see, the data centre cooling bit is…

Deep mind’s protein folding algorithm is probably worth a chunk of change. As far as I know, they’ve been holding onto the secret sauce rather than publishing it.

Re: DeepMind says reinforcement learning is ‘enough’ to reach general AI

#159

Post by jacquesm »

harry8 wrote: post #18

Some Bozo who has heard all this many times before is suspicious of claims from places like Deep Mind who have a financial incentive to make them (keep funding) where there aren't working machines to back that claim up. Some Bozo has no credentials, no reputation, no track record of publications and barely supports the claim they're making with anything much. Some Bozo has no financial incentives or otherwise to opin…

Said anonymous account on HN... If you're going to question other people's credentials, reputation, track record and claims make sure your own are solid. Those who live in glass houses shouldn't throw stones.

Finally, if you're going to attack someone's article: attack the article, not the person that wrote it. This is the lowest level of attack possible: the personal one. It's as ad-hominem as it gets.

Re: DeepMind says reinforcement learning is ‘enough’ to reach general AI

#160

Post by mgoetzke »

zimpenfish wrote: post #94
d110af5ccf wrote: post #78

Earlier quoted context omitted.

Deep Mind is cutting edge ML in general, right? Doesn't Google actively apply the lessons learned all over the place? YouTube content recommendation stands out to me in particular. Translation and automated closed captioning are also obviously ML based. I'd guess that most of the really interesting stuff would be behind the scenes and not immediately visible to end users though.

> YouTube content recommendation stands out to me in particular. If that's "cutting edge ML", then going off my YouTube recommendations, we're back in another AI winter. If I watch one video from a channel I've not seen before, I'll get that channel recommended constantly even if it bears no resemblance to what I normally watch. On my Explore page, the first 22 videos (of which 8 are Fortnite-related!) hold no intere…

Is the Explore page controlled by videos I watched ? Because there isn't a single video on it i would watch. Not one.
Post reply on HN