6 ms·

Original author of the paper here, happy to answer questions. It's a little strange refreshing HN and discovering new popular press about your paper.

cornholio 10y ago
Aside from professional hubris, do you think neuroscience is a good way to learn to create a thinking machine ?
logicallee 10y ago
Here's a [serious] question for you - today by my estimation there are over 1,000 organizations (governments, data warehouses, cloud storage solutions) that use more computing power than it would take to simulate a brain in realtime. So why does ANYBODY think it will NEVER happen? It is as though someone declared, "Powered heavier-than-air flight will never be possible." Uh, of course they will? It's totally obvious? Like how can any reasonable or even poorly-informed, uninvolved person who is in a hurry, possibly not see how completely obvious this is? Help me understand this. Never is a long time. - EDIT: case in point, for some inexplicable reason this comment, a direct question to you stochastician, in the above form (I didn't edit except to add [serious] and the last sentence) was just downvoted twice. (I take it by people who consider everything I just said to be BS.) Just to be clear none of the above is sarcastic, etc. To me - and I suppose you - it's pretty clear what will eventually be happen, and "never" is a strong claim. I completely fail to see how anyone can claim "never" when we're ahead in terms of hardware, today.
- soVeryTired 10y ago
  I think it's partly a problem of not having a roadmap to get from where we are now to where we want to be. Contrast AI with, say, interplanetary space travel. We're a long way away from that too, but there's a more obvious path to getting there. We need better control of nuclear energy, more robust robotics designs, and more efficient environmental control. These are engineering challenges: for the most part, the science is already laid out. With AI the science isn't there. We don't know which direction to move in, let alone how to move in that direction.
  - stochastician 10y ago
    I totally agree with this. In spite of a lot of the recent hype, we still have very limited understanding of intelligence as a science. I agree "never" seems like a strong claim given that we're walking around with 7B existence proofs that it's possible, but for most of us "never" and 300 years are functionally equivalent.
    - logicallee 10y ago
      What makes you think that it would take 300 years, given that humans are being beaten at extremely creative computational tasks like playing Go, already? Machine learning using similar architecture (neural nets) seems to be making astounding progress every day. Why would it take 300 years to make a major breakthrough?
      - soVeryTired 10y ago
        We've made big advances in reinforcement learning and certain types of supervised learning. But most learning is unsupervised. Somehow a human baby bootstraps itself from being a tangle of neurons to sensing, walking, and talking with very little instruction on how to do so.
        logicallee 10y ago
        But monkeys don't start talking (even when surrounded by humans), therefore would you agree that somewhere in our DNA must be the description for how our neurons need to do so? Therefore, that DNA must contain this information? (source code), at some level of abstraction or compressed in some way?
  - logicallee 10y ago
    No!!! There is absolutely, 100% no roadmap for interplanetary space travel, certainly outside the solar system. (Within the solar system is arguable): the 7 planets just discovered are a good example, I love this comment on HN: https://news.ycombinator.com/item?id=13707775 https://news.ycombinator.com/item?id=13707775 "If you had a near perfect hydrogen -> helium fusion engine, it'd take about 6 million tons of fuel (about the mass of the Pyramids of Egypt or 2,000 Saturn V rockets)". Does this sound like a roadmap? "If you had a near perfect hydrogen-> helium fusion engine"... "mass of the Pyramids of Egypt"? Contrast this with the hardware I'm talking about: by my estimation there are 1,000 clusters today, today, that can run the appropriate software if we had it. The idea that someone would call it "impossible" at a time when conventional human thinking challenges such as playing Go, are being mastered by machine using machine learning, seems crazy. I absolutely do not agree with you that we have any roadmap for interplanetary, certainly extrasolar, travel. Contrast this with the Human brain. We can count the neurons (100 billion or so). It uses as much energy as a 20 watt lightbulb[1] but doesn't have light-speed interconnects, unlike server rooms. The roadmap we are following, simulating neural nets on bigger and bigger hardware, is pretty much the exact thing you'd expect us to be doing. It's yielding amazing results in machine learning. Why would anyone think that the biological, slow-firing chemical analog process these huge, super-fast digital processes replicate, will be out of reach literally "forever". It is literally possible that not one human being will ever travel outside the solar system - not in the next 1,000 years; 5,000 years; or ever. The idea that we will "never" simulate a hundred billion analog neurons running at 20 watts seems childishly naive. It seems like saying "no machine will ever fly - period" when machines are already doing more than enough work to do so. Come on. Anyway I am really interested in stochastician's answer here - this is why I wrote them after they offered to answer any questions ("appy to answer questions."). [1] 20 watts
- groby_b 10y ago
  Start by adjusting your estimates. If you assume just neurons and synapses, you're on the petaflop scale. The two first entries on the TOP500 list barely get there. (92 & 38 petaflops, brain assumed at ~30) Supercomputing instances built w/ EC2 are in the 0.25Petaflop range. Assuming that's only a fraction of their machines, maybe the major cloud providers play on the same scale - and at this point, you have to deal with communication across DCs, which might make realtime simulation utterly infeasible. That's 2 yes, and 3-5 maybes, not "over 1000 organizations". Then you're looking at the fact that recent research indicates dendrites carry processing power as well, which would move any brain simulation well into the exascale range. And Moore's law is shakey by now. But given all that is overcome - we seem to be doing OK at the scaling up thing - you would need a complete connectome of the brain. So far, we've succeeded doing that for roundworms. (And 6502's, as the article shows :) The human brain is a far cry off from that. But let's be optimists, and stipulate that is at some point done as well. We're still in the dark, because we have no idea if a fully connected brain will work without being in a specific starting state. We're nowhere even close to map the state of a brain. And at this point, we're definitely in the realm of questions where we don't even know yet if we will find an answer at all. Want more curveballs? There's a non-zero probability that the brain can only be fully explained using QM explanations. You've just significantly shifted the scale of the simulation. I don't know NEVER is an appropriate word, but it's not totally obvious, either. Using the word "obvious" in this context (just like the word "never") means you've reduced the problem to pop-science, where everything is either black or white.
  - logicallee 10y ago
    >If you assume just neurons and synapses, you're on the petaflop scale. Where are you getting this? Did you account for the fact that the average cortical neuron fires 0.16 times per second? [1] Put in terms of gigaherz, that is -- haha just kidding about gigaherz, that is 0.16 herz per neuron. Times 100 billion neurons. "Combined with the observation that 90% of neurons rarely fire" so you actually might not need to touch all 100 billion of them in a cycle. 100 billion * 0.16 gets you literally 16 gigahertz. This is four cores running at 4 gigaherz. Memory can be heavily compressed if necessary, without slipping out of the realtime budget. People don't seem to get that these are ANALOG SIGNALS. These neurons aren't etched in silicon. Simulating that is like simulating a mechanical watch or something - with 100 billion pieces. Your guess is ridiculously too-high. Do you know how long it takes for a signal to cross from one half of the brain to the other? 100 m/s is widely reported progation speed limit, and the average brain length is 167 mm[2] . That means you get 167 mm / 10000 mm = 1.67 milliseconds. That's long enough to cross a server room approximately 25,000 times. (I divided 20meters by C, getting 66 ns - the time at speed of light to cross a 20 meter room, and this value goes into 1.67 ms 25,032 times.) All of my numbers here are super-conservative, above-all the fact that I said only 1,000 server rooms were able to do this today. It's considerably higher. We're ahead of the computational budget required, not behind it. I simply disagree with your calculation. The rest of this paragraph is sarcastic: When mentioning that there's a non-zero chance that neurons use quantum mechanical effects (the way our etched silicon does, since gates are so small now that this has to be accounted for), you forget to also mention that there's a non-zero chance that souls are connected to brains using an as-yet unidentified organ and therefore nothing in the brain can be simulated given any amount of computational power. I mean, the chances of that aren't zero, right? [1] "Based on the energy budget of the brain, it appears that the average cortical neuron fires around 0.16 times per second" - http://aiimpacts.org/rate-of-neuron-firing/ http://aiimpacts.org/rate-of-neuron-firing/ [2] https://www.disabled-world.com/artman/publish/brain-facts.shtml https://www.disabled-world.com/artman/publish/brain-facts.sh...
kevinalexbrown 10y ago
Hi Eric, I haven't read the paper in depth yet, but we've discussed it some as a lab. I love the concept, but I'm a little perplexed by something I hadn't thought of when we first discussed it. It looks like many analyses were performed by binning transistor switching as you might spikes from neurons, or by linearly combining transistor activity across transistors or time. This is slightly puzzling to me, because transistors would appear to not work on the basis of average activity at their level of computational composition. I would be surprised if you could really understand circuit activity at a computational level with most of the measurements you describe. On the other hand, I suspect the methods used work slightly better in a real brain. Of course, the computational purpose of many neurons is difficult to parse or even construct useful hypotheses about. But at the very least, neurons in the outer periphery seem to behave in a way that is at least consistent with firing rate, e.g. impulses at the neuromuscular junction to cause muscle contraction, and retinal ganglion cells to fire at higher rates when the input matches their certain luminance features.
- stochastician 10y ago
  I think that's totally correct! but I'm skeptical that a lot of the non-sensory neurons work on the basis of average activity -- rate codes versus timing codes and all of that. I spent a fair amount of time in the hippocampus where individual spikes. We were always going to run into challenges with how to interpret the voltage levels of the transistor, so we tried to be "fair" and not cheat on either side (we could have decoded the values on the memory bus or the adder numerically, for example). For the next paper, though, we're taking a more careful look at a lot of these encoding questions, as well as trying to use processors with a greater degree of functional specialization so we can more accurately capture what's going on.
nonbel 10y ago
>"Unfortunately, we do not yet have the techniques to also reconstruct the i/o function, neurotransmitter type, ion channel type, I/V curve of each synapse, etc. of each neuron. But even if we did, just as in the case of the processor, we would face the problem of understanding the brain based on its connectome." For the processor I can see the properties of individual transistors being (near) constant, but do you have a reason to think so for neurons in the brain as well?
- AstralStorm 10y ago
  They are very much not due to different chemical environment. Neurons also have built-in crosstalk (synaptic and temporal) which electronics lack.
tmzt 10y ago
What would happen if you evolved a neural network to take the place of a 6502, would it have a similar connectome or morphology?
stochastician 10y ago
Also in the time-honored tradition of HN, the title isn't quite right. 1. we didn't build the connectome, the visual 6502 team did. 2. It's such a good connectome that we can simulate everything perfectly -- a Hansonian "Em", if you will. 3. It's still hard with total visibility to use contemporary techniques to achieve understanding.
- eli_gottlieb 10y ago
  What's the basic point of a connectome? What understanding did anyone expect to achieve from making one? Sure, maybe you can emulate the device, but unless you're a very ardent philosophical connectionist, you're not understanding much. And if you are a very ardent philosophical connectionist, you're claiming that knowing the connections gives us all there is to understand, with no deeper principles than synaptic weights being available. It's as silly as expecting to get a good theory of artificial intelligence by studying artificial neural networks. (I'm aware of the irony in the above statement, but stand by it earnestly. An excellent engineering artifact whose functioning we can't explain is not scientific understanding. An excellent engineering artifact whose functioning we refuse to explain is bad philosophy, too.)
  - stochastician 10y ago
    I think many in the neuroscience community would agree that there's a lot we can learn from connectomes, and there's a lot of value from just having a canonical, authoritative map of the underlying neuroanatomy. Every year there are a few new neuroanatomy papers that say "surprise, area X projects to area Y!" and that's sort of embarrassing given that many people have been studying an area for 30 years and the sudden appearance of dopaminergic projections upsets all their previous models. Also, there exist brain areas and regions where we do in fact have a few good good models, and connectomics has the potential to help us resolve them -- see http://www.nature.com/nature/journal/v500/n7461/full/nature12450.html http://www.nature.com/nature/journal/v500/n7461/full/nature1...
    - apl 10y ago
      Even in these apparently simple feedforward sensory networks, connectomics haven't been the anticipated panacea. There's been a flurry of follow-up papers to Takemura et al., essentially refuting the suggested model. Turns out, even where they should connections don't constrain circuits to a sufficient degree.
georgewfraser 10y ago
The article mentions electron microscopy for connectomics. Do you think electron microscopy or expansion microscopy is more likely to succeed at reconstructing the wiring diagram of real brains?
- stochastician 10y ago
  Expansion microscopy is new and exciting, but also still in progress, and I think struggles to identify some of the traditional areas where you have really fine neuropil. I'm actually most excited about Tony Zador and George Church's DNA-barcode approaches. They will potentially give us different answers, though -- DNA barcodes will give us a graph and detailed information about synapse sites, but not a lot about morphology, for example.

cornholio 10y ago

Aside from professional hubris, do you think neuroscience is a good way to learn to create a thinking machine ?

logicallee 10y ago

Here's a [serious] question for you - today by my estimation there are over 1,000 organizations (governments, data warehouses, cloud storage solutions) that use more computing power than it would take to simulate a brain in realtime. So why does ANYBODY think it will NEVER happen? It is as though someone declared, "Powered heavier-than-air flight will never be possible." Uh, of course they will? It's totally obvious? Like how can any reasonable or even poorly-informed, uninvolved person who is in a hurry, possibly not see how completely obvious this is? Help me understand this. Never is a long time. - EDIT: case in point, for some inexplicable reason this comment, a direct question to you stochastician, in the above form (I didn't edit except to add [serious] and the last sentence) was just downvoted twice. (I take it by people who consider everything I just said to be BS.) Just to be clear none of the above is sarcastic, etc. To me - and I suppose you - it's pretty clear what will eventually be happen, and "never" is a strong claim. I completely fail to see how anyone can claim "never" when we're ahead in terms of hardware, today.

soVeryTired 10y ago
I think it's partly a problem of not having a roadmap to get from where we are now to where we want to be. Contrast AI with, say, interplanetary space travel. We're a long way away from that too, but there's a more obvious path to getting there. We need better control of nuclear energy, more robust robotics designs, and more efficient environmental control. These are engineering challenges: for the most part, the science is already laid out. With AI the science isn't there. We don't know which direction to move in, let alone how to move in that direction.
- stochastician 10y ago
  I totally agree with this. In spite of a lot of the recent hype, we still have very limited understanding of intelligence as a science. I agree "never" seems like a strong claim given that we're walking around with 7B existence proofs that it's possible, but for most of us "never" and 300 years are functionally equivalent.
  - logicallee 10y ago
    What makes you think that it would take 300 years, given that humans are being beaten at extremely creative computational tasks like playing Go, already? Machine learning using similar architecture (neural nets) seems to be making astounding progress every day. Why would it take 300 years to make a major breakthrough?
    - soVeryTired 10y ago
      We've made big advances in reinforcement learning and certain types of supervised learning. But most learning is unsupervised. Somehow a human baby bootstraps itself from being a tangle of neurons to sensing, walking, and talking with very little instruction on how to do so.
      - logicallee 10y ago
        But monkeys don't start talking (even when surrounded by humans), therefore would you agree that somewhere in our DNA must be the description for how our neurons need to do so? Therefore, that DNA must contain this information? (source code), at some level of abstraction or compressed in some way?
- logicallee 10y ago
  No!!! There is absolutely, 100% no roadmap for interplanetary space travel, certainly outside the solar system. (Within the solar system is arguable): the 7 planets just discovered are a good example, I love this comment on HN: https://news.ycombinator.com/item?id=13707775 https://news.ycombinator.com/item?id=13707775 "If you had a near perfect hydrogen -> helium fusion engine, it'd take about 6 million tons of fuel (about the mass of the Pyramids of Egypt or 2,000 Saturn V rockets)". Does this sound like a roadmap? "If you had a near perfect hydrogen-> helium fusion engine"... "mass of the Pyramids of Egypt"? Contrast this with the hardware I'm talking about: by my estimation there are 1,000 clusters today, today, that can run the appropriate software if we had it. The idea that someone would call it "impossible" at a time when conventional human thinking challenges such as playing Go, are being mastered by machine using machine learning, seems crazy. I absolutely do not agree with you that we have any roadmap for interplanetary, certainly extrasolar, travel. Contrast this with the Human brain. We can count the neurons (100 billion or so). It uses as much energy as a 20 watt lightbulb[1] but doesn't have light-speed interconnects, unlike server rooms. The roadmap we are following, simulating neural nets on bigger and bigger hardware, is pretty much the exact thing you'd expect us to be doing. It's yielding amazing results in machine learning. Why would anyone think that the biological, slow-firing chemical analog process these huge, super-fast digital processes replicate, will be out of reach literally "forever". It is literally possible that not one human being will ever travel outside the solar system - not in the next 1,000 years; 5,000 years; or ever. The idea that we will "never" simulate a hundred billion analog neurons running at 20 watts seems childishly naive. It seems like saying "no machine will ever fly - period" when machines are already doing more than enough work to do so. Come on. Anyway I am really interested in stochastician's answer here - this is why I wrote them after they offered to answer any questions ("appy to answer questions."). [1] 20 watts
groby_b 10y ago
Start by adjusting your estimates. If you assume just neurons and synapses, you're on the petaflop scale. The two first entries on the TOP500 list barely get there. (92 & 38 petaflops, brain assumed at ~30) Supercomputing instances built w/ EC2 are in the 0.25Petaflop range. Assuming that's only a fraction of their machines, maybe the major cloud providers play on the same scale - and at this point, you have to deal with communication across DCs, which might make realtime simulation utterly infeasible. That's 2 yes, and 3-5 maybes, not "over 1000 organizations". Then you're looking at the fact that recent research indicates dendrites carry processing power as well, which would move any brain simulation well into the exascale range. And Moore's law is shakey by now. But given all that is overcome - we seem to be doing OK at the scaling up thing - you would need a complete connectome of the brain. So far, we've succeeded doing that for roundworms. (And 6502's, as the article shows :) The human brain is a far cry off from that. But let's be optimists, and stipulate that is at some point done as well. We're still in the dark, because we have no idea if a fully connected brain will work without being in a specific starting state. We're nowhere even close to map the state of a brain. And at this point, we're definitely in the realm of questions where we don't even know yet if we will find an answer at all. Want more curveballs? There's a non-zero probability that the brain can only be fully explained using QM explanations. You've just significantly shifted the scale of the simulation. I don't know NEVER is an appropriate word, but it's not totally obvious, either. Using the word "obvious" in this context (just like the word "never") means you've reduced the problem to pop-science, where everything is either black or white.
- logicallee 10y ago
  >If you assume just neurons and synapses, you're on the petaflop scale. Where are you getting this? Did you account for the fact that the average cortical neuron fires 0.16 times per second? [1] Put in terms of gigaherz, that is -- haha just kidding about gigaherz, that is 0.16 herz per neuron. Times 100 billion neurons. "Combined with the observation that 90% of neurons rarely fire" so you actually might not need to touch all 100 billion of them in a cycle. 100 billion * 0.16 gets you literally 16 gigahertz. This is four cores running at 4 gigaherz. Memory can be heavily compressed if necessary, without slipping out of the realtime budget. People don't seem to get that these are ANALOG SIGNALS. These neurons aren't etched in silicon. Simulating that is like simulating a mechanical watch or something - with 100 billion pieces. Your guess is ridiculously too-high. Do you know how long it takes for a signal to cross from one half of the brain to the other? 100 m/s is widely reported progation speed limit, and the average brain length is 167 mm[2] . That means you get 167 mm / 10000 mm = 1.67 milliseconds. That's long enough to cross a server room approximately 25,000 times. (I divided 20meters by C, getting 66 ns - the time at speed of light to cross a 20 meter room, and this value goes into 1.67 ms 25,032 times.) All of my numbers here are super-conservative, above-all the fact that I said only 1,000 server rooms were able to do this today. It's considerably higher. We're ahead of the computational budget required, not behind it. I simply disagree with your calculation. The rest of this paragraph is sarcastic: When mentioning that there's a non-zero chance that neurons use quantum mechanical effects (the way our etched silicon does, since gates are so small now that this has to be accounted for), you forget to also mention that there's a non-zero chance that souls are connected to brains using an as-yet unidentified organ and therefore nothing in the brain can be simulated given any amount of computational power. I mean, the chances of that aren't zero, right? [1] "Based on the energy budget of the brain, it appears that the average cortical neuron fires around 0.16 times per second" - http://aiimpacts.org/rate-of-neuron-firing/ http://aiimpacts.org/rate-of-neuron-firing/ [2] https://www.disabled-world.com/artman/publish/brain-facts.shtml https://www.disabled-world.com/artman/publish/brain-facts.sh...

kevinalexbrown 10y ago

Hi Eric, I haven't read the paper in depth yet, but we've discussed it some as a lab. I love the concept, but I'm a little perplexed by something I hadn't thought of when we first discussed it. It looks like many analyses were performed by binning transistor switching as you might spikes from neurons, or by linearly combining transistor activity across transistors or time. This is slightly puzzling to me, because transistors would appear to not work on the basis of average activity at their level of computational composition. I would be surprised if you could really understand circuit activity at a computational level with most of the measurements you describe. On the other hand, I suspect the methods used work slightly better in a real brain. Of course, the computational purpose of many neurons is difficult to parse or even construct useful hypotheses about. But at the very least, neurons in the outer periphery seem to behave in a way that is at least consistent with firing rate, e.g. impulses at the neuromuscular junction to cause muscle contraction, and retinal ganglion cells to fire at higher rates when the input matches their certain luminance features.

stochastician 10y ago
I think that's totally correct! but I'm skeptical that a lot of the non-sensory neurons work on the basis of average activity -- rate codes versus timing codes and all of that. I spent a fair amount of time in the hippocampus where individual spikes. We were always going to run into challenges with how to interpret the voltage levels of the transistor, so we tried to be "fair" and not cheat on either side (we could have decoded the values on the memory bus or the adder numerically, for example). For the next paper, though, we're taking a more careful look at a lot of these encoding questions, as well as trying to use processors with a greater degree of functional specialization so we can more accurately capture what's going on.

nonbel 10y ago

>"Unfortunately, we do not yet have the techniques to also reconstruct the i/o function, neurotransmitter type, ion channel type, I/V curve of each synapse, etc. of each neuron. But even if we did, just as in the case of the processor, we would face the problem of understanding the brain based on its connectome." For the processor I can see the properties of individual transistors being (near) constant, but do you have a reason to think so for neurons in the brain as well?

AstralStorm 10y ago
They are very much not due to different chemical environment. Neurons also have built-in crosstalk (synaptic and temporal) which electronics lack.

tmzt 10y ago

What would happen if you evolved a neural network to take the place of a 6502, would it have a similar connectome or morphology?

stochastician 10y ago

Also in the time-honored tradition of HN, the title isn't quite right. 1. we didn't build the connectome, the visual 6502 team did. 2. It's such a good connectome that we can simulate everything perfectly -- a Hansonian "Em", if you will. 3. It's still hard with total visibility to use contemporary techniques to achieve understanding.

eli_gottlieb 10y ago
What's the basic point of a connectome? What understanding did anyone expect to achieve from making one? Sure, maybe you can emulate the device, but unless you're a very ardent philosophical connectionist, you're not understanding much. And if you are a very ardent philosophical connectionist, you're claiming that knowing the connections gives us all there is to understand, with no deeper principles than synaptic weights being available. It's as silly as expecting to get a good theory of artificial intelligence by studying artificial neural networks. (I'm aware of the irony in the above statement, but stand by it earnestly. An excellent engineering artifact whose functioning we can't explain is not scientific understanding. An excellent engineering artifact whose functioning we refuse to explain is bad philosophy, too.)
- stochastician 10y ago
  I think many in the neuroscience community would agree that there's a lot we can learn from connectomes, and there's a lot of value from just having a canonical, authoritative map of the underlying neuroanatomy. Every year there are a few new neuroanatomy papers that say "surprise, area X projects to area Y!" and that's sort of embarrassing given that many people have been studying an area for 30 years and the sudden appearance of dopaminergic projections upsets all their previous models. Also, there exist brain areas and regions where we do in fact have a few good good models, and connectomics has the potential to help us resolve them -- see http://www.nature.com/nature/journal/v500/n7461/full/nature12450.html http://www.nature.com/nature/journal/v500/n7461/full/nature1...
  - apl 10y ago
    Even in these apparently simple feedforward sensory networks, connectomics haven't been the anticipated panacea. There's been a flurry of follow-up papers to Takemura et al., essentially refuting the suggested model. Turns out, even where they should connections don't constrain circuits to a sufficient degree.

georgewfraser 10y ago

The article mentions electron microscopy for connectomics. Do you think electron microscopy or expansion microscopy is more likely to succeed at reconstructing the wiring diagram of real brains?

stochastician 10y ago
Expansion microscopy is new and exciting, but also still in progress, and I think struggles to identify some of the traditional areas where you have really fine neuropil. I'm actually most excited about Tony Zador and George Church's DNA-barcode approaches. They will potentially give us different answers, though -- DNA barcodes will give us a graph and detailed information about synapse sites, but not a lot about morphology, for example.