4 ms·

There are many thing that science doesn't yet know conclusively about neural architectures. Arguing that a machine such as described in the article will get us

by geuis 14y ago

There are many thing that science doesn't yet know conclusively about neural architectures. Arguing that a machine such as described in the article will get us nowhere massively misses the point. We live in the age of simulation in science. Computation has reached the point where, in many disparate systems such as weather models, particle physics, protein folding, and neural architecture, building complex simulations lets investigators freely play with hypothesis and compare the results to real-world data. Hypothesis are proved or disproved, science advances, and we know that much more about how such complex systems indeed work. The author of this article derides this approach, saying that we just don't have enough information to do anything useful with the machine. I call bullshit. One of the entire points of the machine is to allow experimenters perform experiments that are either impossible or morally unethical on biological creatures. In a few years, suppose a research decides to change the algorithm about how simulated neurons exchange a neurotransmitter or an electrical pulse. The simulation then produces patterns very similar or identical to an FMRI or electrocardiogram of a living brain. Biologists can then use that as an unexpected avenue to exist living neurons to see if they react in the same way. Its an instance of simulation allowing experimentation that illuminates the real deal.

Xcelerate 14y ago
> ...the machine is to allow experimenters to perform experiments that are either impossible or morally unethical on biological creatures Ignoring the fact that whether this ever happens is up for debate, why would it be moral to torture a simulated animal as long as you're not torturing a "real" one? > In a few years, suppose a research decides... You vastly overestimate what research will produce in a few years.
- javert 14y ago
  why would it be moral to torture a simulated animal as long as you're not torturing a "real" one? Morality is a human concept that should serve human ends--individually and socially. As such, performing experiments on a dog to gain more knowledge is moral.
  - jessedhillon 14y ago
    Morality is a human concept that should serve human ends--individually and socially. This is a completely unsubstantiated assumption. As such, performing experiments on a dog to gain more knowledge is moral. This relies on the unjustified assumption that causing suffering of a non-human is not immoral or amoral.
    - javert 14y ago
      Not completely unsubstantiated, but I'm not going to sit here and teach philosophy (mainly because nobody would listen and I wouldn't get paid for doing it). This relies on the unjustified assumption that causing suffering of a non-human is not immoral or amoral. That's like saying "God must exist, until you prove that he doesn't." In other words, on that particular point, you're the one making a positive claim.
      - jessedhillon 14y ago
        Not exactly. You have a particular answer to the questions What is morality, and What is moral, and have asserted those answers as true. It is not necessarily the default state of the Universe that some/all actions are immoral, or amoral -- I am contesting your assertion that this action is without moral consequence. Edit: actually, you claimed without justification to know that the action was moral. In general, I am highly suspect of anyone who claims to know with certainty what morality is or isn't, or why it exists (or doesn't). People have been contemplating this for hundreds of years. It is very unlikely that you (or anyone) have discovered a new, empirically verifiable truth about morality. It is more likely that such claims are overconfident and unjustified.
        javert 14y ago
        I was saying, "presume no moral consequence in a particular case until it's show otherwise," and you're saying "presume moral consequence in a particular case until it's shown otherwise." Or maybe you're saying, "don't presume either way until you have evidence." I think that at the end of the day, that third option is actually the right one. So, rather than saying, "Hey, you're the one making a positive claim," I should have just said, "Actually, here is what is right, and here's why." Of course, the "why" would be limited in scope, because I'd rather direct you to external information that I agree with than write a philosophical treatise in the comments of HN. Anyway, I guess at this point I'm talking about how to talk, which is starting to get "too meta." Forgive me. To move on to something more interesting: In general, I am highly suspect of anyone who claims to know with certainty what morality is or isn't, or why it exists (or doesn't). People have been contemplating this for hundreds of years. It is very unlikely that you (or anyone) have discovered a new, empirically verifiable truth about morality. It is more likely that such claims are overconfident and unjustified. I basically agree with your analysis, here. But, unlikely things actually do happen. In this case, I believe they have. I have some info on my profile that will direct you to more information if you are interested.
- geuis 14y ago
  Your first point is valid and the answer can be debated both ways. From my point of view, scientific advancements through the centuries have come at the cost of what we consider now immoral treatment of animals and humans. I feel that we're going to see a repeat of some of this behavior with simulated entities for a while once we gain the capability. It may be accepted, and not even an issue for the public at large, until some signs are being found that the entities are conscious. Hopefully we collectively approach this area of research with the foreknowledge not to repeat previous mistakes. In the end, I feel strongly that our moral definitions will expand to include simulated entities. This is vital for the existence of AI-type minds (constructed and uploaded) to co-exist in society with flesh and blood humans. A further note to that effect is the recent movement to give sentience-rights to cetaceans and the protection that goes with it. Its encouraging to see that movement taking off, as it will bode well for our hypothetical simulated fellows in the future. To your second point, I believe you vastly under estimate the research that will be done in the next few years. There are incredible things already done in labs around the world that aren't widely viewed yet. In the next few years that research is going to start exiting the lab and start hitting the marketplace. The last 20 years of advancement are nothing compared to the next 20. We would all do well to bet on a stranger, faster future rather than one that seems like the status quo of the moment.
  - Xcelerate 14y ago
    > To your second point, I believe you vastly under estimate the research that will be done in the next few years. Odd. As my research is in simulation using massively parallel supercomputers, I'd suggest that unless there are (secretive) private research groups out there that are way ahead of universities and have access to pflops level computing, I don't think we'll be seeing accurate simulated FMRIs any time soon.
new299 14y ago
What simulations often lack is any clear validation against reality. The true test of a simulation is it's ability to predict the behaviour of a real system, it's not enough to say "their construction has some similar features so it's probably ok". Weather simulations are generally reasonable, though limited in this regard. Other scientific simulations tend to be a little more hazy, the main aim of course being to produce a publication. You said "Hypothesis are proved or disproved, science advances, and we know that much more about how such complex systems indeed work". Could you provide some examples? In my experience I've rarely seen simulations have have proved or disproved hypothesis. This is an honest question, if you have some references to significant results from simulation I'd be interested in reading them.
nsp 14y ago
MMijsexfcr

by geuis 14y ago

Xcelerate 14y ago

> ...the machine is to allow experimenters to perform experiments that are either impossible or morally unethical on biological creatures Ignoring the fact that whether this ever happens is up for debate, why would it be moral to torture a simulated animal as long as you're not torturing a "real" one? > In a few years, suppose a research decides... You vastly overestimate what research will produce in a few years.

javert 14y ago
why would it be moral to torture a simulated animal as long as you're not torturing a "real" one? Morality is a human concept that should serve human ends--individually and socially. As such, performing experiments on a dog to gain more knowledge is moral.
- jessedhillon 14y ago
  Morality is a human concept that should serve human ends--individually and socially. This is a completely unsubstantiated assumption. As such, performing experiments on a dog to gain more knowledge is moral. This relies on the unjustified assumption that causing suffering of a non-human is not immoral or amoral.
  - javert 14y ago
    Not completely unsubstantiated, but I'm not going to sit here and teach philosophy (mainly because nobody would listen and I wouldn't get paid for doing it). This relies on the unjustified assumption that causing suffering of a non-human is not immoral or amoral. That's like saying "God must exist, until you prove that he doesn't." In other words, on that particular point, you're the one making a positive claim.
    - jessedhillon 14y ago
      Not exactly. You have a particular answer to the questions What is morality, and What is moral, and have asserted those answers as true. It is not necessarily the default state of the Universe that some/all actions are immoral, or amoral -- I am contesting your assertion that this action is without moral consequence. Edit: actually, you claimed without justification to know that the action was moral. In general, I am highly suspect of anyone who claims to know with certainty what morality is or isn't, or why it exists (or doesn't). People have been contemplating this for hundreds of years. It is very unlikely that you (or anyone) have discovered a new, empirically verifiable truth about morality. It is more likely that such claims are overconfident and unjustified.
      - javert 14y ago
        I was saying, "presume no moral consequence in a particular case until it's show otherwise," and you're saying "presume moral consequence in a particular case until it's shown otherwise." Or maybe you're saying, "don't presume either way until you have evidence." I think that at the end of the day, that third option is actually the right one. So, rather than saying, "Hey, you're the one making a positive claim," I should have just said, "Actually, here is what is right, and here's why." Of course, the "why" would be limited in scope, because I'd rather direct you to external information that I agree with than write a philosophical treatise in the comments of HN. Anyway, I guess at this point I'm talking about how to talk, which is starting to get "too meta." Forgive me. To move on to something more interesting: In general, I am highly suspect of anyone who claims to know with certainty what morality is or isn't, or why it exists (or doesn't). People have been contemplating this for hundreds of years. It is very unlikely that you (or anyone) have discovered a new, empirically verifiable truth about morality. It is more likely that such claims are overconfident and unjustified. I basically agree with your analysis, here. But, unlikely things actually do happen. In this case, I believe they have. I have some info on my profile that will direct you to more information if you are interested.
geuis 14y ago
Your first point is valid and the answer can be debated both ways. From my point of view, scientific advancements through the centuries have come at the cost of what we consider now immoral treatment of animals and humans. I feel that we're going to see a repeat of some of this behavior with simulated entities for a while once we gain the capability. It may be accepted, and not even an issue for the public at large, until some signs are being found that the entities are conscious. Hopefully we collectively approach this area of research with the foreknowledge not to repeat previous mistakes. In the end, I feel strongly that our moral definitions will expand to include simulated entities. This is vital for the existence of AI-type minds (constructed and uploaded) to co-exist in society with flesh and blood humans. A further note to that effect is the recent movement to give sentience-rights to cetaceans and the protection that goes with it. Its encouraging to see that movement taking off, as it will bode well for our hypothetical simulated fellows in the future. To your second point, I believe you vastly under estimate the research that will be done in the next few years. There are incredible things already done in labs around the world that aren't widely viewed yet. In the next few years that research is going to start exiting the lab and start hitting the marketplace. The last 20 years of advancement are nothing compared to the next 20. We would all do well to bet on a stranger, faster future rather than one that seems like the status quo of the moment.
- Xcelerate 14y ago
  > To your second point, I believe you vastly under estimate the research that will be done in the next few years. Odd. As my research is in simulation using massively parallel supercomputers, I'd suggest that unless there are (secretive) private research groups out there that are way ahead of universities and have access to pflops level computing, I don't think we'll be seeing accurate simulated FMRIs any time soon.

new299 14y ago

What simulations often lack is any clear validation against reality. The true test of a simulation is it's ability to predict the behaviour of a real system, it's not enough to say "their construction has some similar features so it's probably ok". Weather simulations are generally reasonable, though limited in this regard. Other scientific simulations tend to be a little more hazy, the main aim of course being to produce a publication. You said "Hypothesis are proved or disproved, science advances, and we know that much more about how such complex systems indeed work". Could you provide some examples? In my experience I've rarely seen simulations have have proved or disproved hypothesis. This is an honest question, if you have some references to significant results from simulation I'd be interested in reading them.

nsp 14y ago

MMijsexfcr