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

101–110 of 312 posts

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

#101

Post by rich_sasha »

d110af5ccf wrote: post #78
rich_sasha wrote: post #64

Earlier quoted context omitted.

Is Deep Mind a “booming business”? They are achieving great things academically, but their business successes are either kept secret or mostly absent. All I know about is the Google data centre cooling scheduling, probably a big saving for Google but hardly an achievement that in its own professes their business success.

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 the only palpable (public) business success.

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

#102

Post by rich_sasha »

Clewza313 wrote: post #62
londons_explore wrote: post #58

Earlier quoted context omitted.

Nature has a massive incentive to make good use of energy from light through photosynthesis. Billions of plants compete, and whoever can get most out of the sun will win out. Yet manmade solar cells are more efficient by nearly all measures.

Except that manmade solar cells are pretty bad at repairing or replicating themselves.

Or growing out of literally nothing but dirt and water.

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

#103

Post by hestefisk »

For those of you ‘generalised’ AI sceptics like me, I can highly recommend reading Søren Brier’s book on Cybersemiotics (“why information is not enough”). It’s a comprehensive reader into the physicalistic, reductionist field of AI and all of its shortcomings. General AI implies the ability to abduct (not just deduction and induction), which I highly doubt will ever be possible.

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

#105

Post by kakadzhun »

Even if Reinforcement Learning is 'enough', it will be held back by whatever methods used to implement it (e.g. Deep Neural Networks). As the algorithms get more advanced, from some point onwards, to build the general AI, you first need a general AI to tell you the correct hyperparameters so that the pile of methods would work well in tandem.

Last but not least, you will be bound by your inability to accurately communicate exactly the behaviour you want out of the AI because you are incapable of writing down a mathematical function that would induce the behaviour in the learner. And then you wonder why the general AI decided to pull the plug on your grandma and try to use that resource for something else instead.

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

#106

Post by funklute »

hestefisk wrote: post #103

For those of you ‘generalised’ AI sceptics like me, I can highly recommend reading Søren Brier’s book on Cybersemiotics (“why information is not enough”). It’s a comprehensive reader into the physicalistic, reductionist field of AI and all of its shortcomings. General AI implies the ability to abduct (not just deduction and induction), which I highly doubt will ever be possible.

> which I highly doubt will ever be possible

What prevents an AI from performing abduction? While I've never thought about it before, it intuitively seems like a pretty straightforward thing to implement....

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

#107

Post by 666lumberjack »

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…

How often do you use YouTube? Personally I am a very heavy user and in my experience the obsession with a new video kind you watch only lasts for a few recommendations unless you lean into it.

I would guess about two thirds of the channels I consistently watch I originally discovered through algorithm recommendations. I think it works extremely well.

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

#108

Post by throwaways885 »

d110af5ccf wrote: post #78
rich_sasha wrote: post #64

Earlier quoted context omitted.

Is Deep Mind a “booming business”? They are achieving great things academically, but their business successes are either kept secret or mostly absent. All I know about is the Google data centre cooling scheduling, probably a big saving for Google but hardly an achievement that in its own professes their business success.

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.

They have an applied division which applies ML to Google products. I suspect they are very valuable in $ terms just for the work listed here: https://deepmind.com/impact. Google's entire business from the start was doing research and bringing it to the masses, so this shouldn't really surprise anyone.

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

#109

Post by reedf1 »

hestefisk wrote: post #103

For those of you ‘generalised’ AI sceptics like me, I can highly recommend reading Søren Brier’s book on Cybersemiotics (“why information is not enough”). It’s a comprehensive reader into the physicalistic, reductionist field of AI and all of its shortcomings. General AI implies the ability to abduct (not just deduction and induction), which I highly doubt will ever be possible.

How does he resolve that humans achieve the ability to "abduct"? I just don't buy that there is something ineffable about humans - and even if there is, why can't we just plug that process into a computer?

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

#110

Post by rich_sasha »

Basically, any problem with a solution fits into RL: reward of 1 if you are AGI and 0 otherwise. Go learn.

This setting on its own is meaningless! The “how” of the RL agent is not even 99% of the problem, it is all of it.

Given our understanding of both DL and neuroscience, it is not even clear to me that we can say with confidence that Neural Networks are a sufficiently expressive architecture to cover an AGI.

The human brain is a deep net, sort of, but there is also plenty going on in our brains that we don’t understand. It could be that the magic sprinkle is orthogonal to DL and we just don’t know about it yet.

Post reply on HN