3 ms·

I co-authored a paper exploring this topic few years ago, while I was pretty excited about the possiblity of using embeddings for generalization. "Towards conc

I co-authored a paper exploring this topic few years ago, while I was pretty excited about the possiblity of using embeddings for generalization. "Towards conceptual generalization in the embedding space" https://arxiv.org/abs/1906.01873 https://arxiv.org/abs/1906.01873 I still think the approach outlined in the paper (using embeddings to map the physical world) is sound especially for the field of self-driving which is in dire need of generalization, but I've since changed my mind and currently do not believe we can achieve AGI (ever). While embeddings are a great tool for compressing information, they do not provide inherent mechanisms for manipulating the information stored in order to generalize and infer outcomes in new, unseen situations. And even if we would start producing embeddings in a way where they would have some basic understanding of the physical world, we could never achieve it to the level of detail necessary - because physical world is not a discrete function. Otherwise we would be creating a perfect simulation (within a simulation?). And the last time I was playing God, was in "Populous".

ascar 5y ago
> I've since changed my mind and currently do not believe we can achieve AGI (ever). Considering we (as in humans) developed general intelligence, isn't that already in contradiction with your statement? If it happened for us and is "easily" replicated through our DNA, it certainly can be developed again in an artificial medium. But the solution might not have anything to do with what we call machine learning today and sure we might go extinct before (but I didn't have the feeling that's what you were implying).
- trompetenaccoun 5y ago
  It's semantics at this point but we did not create ourselves, it was a complex process that took billions of years to create each one of us. Something being conceivable isn't the same as it being practically possible. I can imagine what you propose, but the same goes for traveling to distant stars or a time machine for going to the future. All perfectly possible in theory.
  - jamiek88 5y ago
    Yeah but interstellar and time travel haven’t been done, or at least we haven’t observed such. Flight had. Intelligence has.
    - trompetenaccoun 5y ago
      Intelligence is an abstract concept, it depends on what exactly one means by that. I have watched a rockets take off from earth. I have never seen a self-aware machine.
    - inter_netuser 5y ago
      What are you talking about? We have interstellar objects intrude and fly by our solar system all the time, on several occasions.
- freediver 5y ago
  It is not a contradiction as I meant "achieving" in the context of creating it (through software). The fact it happened to us is undeniable (from our perspective), but the how/why of it is still one of the many mysteries of the universe - one we will likely never solve.
  - extheat 5y ago
    I’m curious why you think that. Do you think it’s a fundamental problem with the discrete nature of traditional computers? Or a problem with scale and computational limits? If it’s the latter, if a hypothetical computer has unlimited time and memory capacity, why do you think AGI would remain impossible?
    - freediver 5y ago
      Machines are good at computation, which is not equal to reasoning, but rather a subset of reasoning. And not only they are good at computation, but they are exceptionally good at it - I have no illusion of trying to compete with a machine doing square roots or playing chess. And increasingly harder problems are being expressed as computation problems, with more or less success - most famously probably self-driving. But at the end of the day it feels like using an increasingly longer ladder to reach the surface of the Moon. While imaginable, and every time we extend the ladder the Moon does get closer, it is fundamentaly impossible.
      - jameshart 5y ago
        Ever since Gödel we’ve had a pretty convincing proof that there is nothing that you can do in terms of reasoning that can’t be expressed using computation. And since Turing we’ve got a framework that shows there’s nothing computable that you can’t compute using a universal computer. So unless there’s something mystical beyond the realm of mathematics to ‘reasoning’ it can’t be a superset of computing. If a finite amount of matter in a brain with a finite amount of energy can do it, then a universal computing machine with a finite amount of storage and a finite amount of time can do it.
staticassertion 5y ago
> currently do not believe we can achieve AGI (ever). Do you mean with embeddings as the approach, or in general?
- freediver 5y ago
  I think AGI will remain out of reach. Even a simpler thing like level 5 self-driving, which is only like 0.3 AGI or something, will remain forever out of reach no matter how much compute we throw at it (though I also think that if we ever reach 0.3 AGI we will also reach 100%). The reason is that the mundane world keeps surprising us everywhere we look and constantly keeps creating more questions than answers. Just look at the questions field of quantum mechanics is trying to tackle, but also every other field of research science - astronomy, genetics, biology, antropology even mathematics... Now imagine trying to keep up with all that - by writing code. Also, mastering these things would make us 'God'.
axg11 5y ago
Thanks for your perspective. We’re still in disagreement but I wouldn’t bet on either side of the AGI debate with any significant conviction. Embeddings are very good at a few things: combining concepts (addition), untangling commonalities (subtraction) and determining similarity between concepts (distance). > While embeddings are a great tool for compressing information, they do not provide inherent mechanisms for manipulating the information stored What are the manipulations you’re referring to? I would love to learn more. From my understanding, embeddings actually provide great generalisation. If you have a well conditioned embedding space then you can interpolate into previously unseen parts of that space and still get sensible results. That is generalisation to me. Many current ML methods do _not_ result in a fully meaningful embedding space but my hunch is that we will get there with future insights and advances.
- freediver 5y ago
  > We’re still in disagreement but I wouldn’t bet on either side of the AGI debate with any significant conviction. That is probably a superior position to hold. I am agnostic by nature, and interestingly this is one of the rare topics I've taken a hard position on. It could be a result of the years spent in the field but also some kind of bias. > What are the manipulations you’re referring to? Need to take a step back and mention that in the field of AI there is a great debate between symbolic and non-symbolic approaches. (and after decades spent with AI under symbolic approaches domination we are now in the golden age of non-symbolic AI; with symbolic starting to have a comeback. this podcast can be a good starting point to learn more https://lexfridman.com/gary-marcus/ https://lexfridman.com/gary-marcus/ - although I disagree with GM on many things - and this tweet for learning about symbolic making a comeback https://twitter.com/hardmaru/status/1470847417193209856 https://twitter.com/hardmaru/status/1470847417193209856) Basically embeddings are "non-symbolic AI" (which is great and this is where their huge potential stems from), but the very way they are generated and then later utlized is completely "symbolic". Which means the the limits of embeddings is defined by the limits of (in this case human written) symbols used to define them. Hope that makes sense.

ascar 5y ago

> I've since changed my mind and currently do not believe we can achieve AGI (ever). Considering we (as in humans) developed general intelligence, isn't that already in contradiction with your statement? If it happened for us and is "easily" replicated through our DNA, it certainly can be developed again in an artificial medium. But the solution might not have anything to do with what we call machine learning today and sure we might go extinct before (but I didn't have the feeling that's what you were implying).

trompetenaccoun 5y ago
It's semantics at this point but we did not create ourselves, it was a complex process that took billions of years to create each one of us. Something being conceivable isn't the same as it being practically possible. I can imagine what you propose, but the same goes for traveling to distant stars or a time machine for going to the future. All perfectly possible in theory.
- jamiek88 5y ago
  Yeah but interstellar and time travel haven’t been done, or at least we haven’t observed such. Flight had. Intelligence has.
  - trompetenaccoun 5y ago
    Intelligence is an abstract concept, it depends on what exactly one means by that. I have watched a rockets take off from earth. I have never seen a self-aware machine.
  - inter_netuser 5y ago
    What are you talking about? We have interstellar objects intrude and fly by our solar system all the time, on several occasions.
freediver 5y ago
It is not a contradiction as I meant "achieving" in the context of creating it (through software). The fact it happened to us is undeniable (from our perspective), but the how/why of it is still one of the many mysteries of the universe - one we will likely never solve.
- extheat 5y ago
  I’m curious why you think that. Do you think it’s a fundamental problem with the discrete nature of traditional computers? Or a problem with scale and computational limits? If it’s the latter, if a hypothetical computer has unlimited time and memory capacity, why do you think AGI would remain impossible?
  - freediver 5y ago
    Machines are good at computation, which is not equal to reasoning, but rather a subset of reasoning. And not only they are good at computation, but they are exceptionally good at it - I have no illusion of trying to compete with a machine doing square roots or playing chess. And increasingly harder problems are being expressed as computation problems, with more or less success - most famously probably self-driving. But at the end of the day it feels like using an increasingly longer ladder to reach the surface of the Moon. While imaginable, and every time we extend the ladder the Moon does get closer, it is fundamentaly impossible.
    - jameshart 5y ago
      Ever since Gödel we’ve had a pretty convincing proof that there is nothing that you can do in terms of reasoning that can’t be expressed using computation. And since Turing we’ve got a framework that shows there’s nothing computable that you can’t compute using a universal computer. So unless there’s something mystical beyond the realm of mathematics to ‘reasoning’ it can’t be a superset of computing. If a finite amount of matter in a brain with a finite amount of energy can do it, then a universal computing machine with a finite amount of storage and a finite amount of time can do it.

staticassertion 5y ago

> currently do not believe we can achieve AGI (ever). Do you mean with embeddings as the approach, or in general?

freediver 5y ago
I think AGI will remain out of reach. Even a simpler thing like level 5 self-driving, which is only like 0.3 AGI or something, will remain forever out of reach no matter how much compute we throw at it (though I also think that if we ever reach 0.3 AGI we will also reach 100%). The reason is that the mundane world keeps surprising us everywhere we look and constantly keeps creating more questions than answers. Just look at the questions field of quantum mechanics is trying to tackle, but also every other field of research science - astronomy, genetics, biology, antropology even mathematics... Now imagine trying to keep up with all that - by writing code. Also, mastering these things would make us 'God'.

axg11 5y ago

Thanks for your perspective. We’re still in disagreement but I wouldn’t bet on either side of the AGI debate with any significant conviction. Embeddings are very good at a few things: combining concepts (addition), untangling commonalities (subtraction) and determining similarity between concepts (distance). > While embeddings are a great tool for compressing information, they do not provide inherent mechanisms for manipulating the information stored What are the manipulations you’re referring to? I would love to learn more. From my understanding, embeddings actually provide great generalisation. If you have a well conditioned embedding space then you can interpolate into previously unseen parts of that space and still get sensible results. That is generalisation to me. Many current ML methods do _not_ result in a fully meaningful embedding space but my hunch is that we will get there with future insights and advances.

freediver 5y ago
> We’re still in disagreement but I wouldn’t bet on either side of the AGI debate with any significant conviction. That is probably a superior position to hold. I am agnostic by nature, and interestingly this is one of the rare topics I've taken a hard position on. It could be a result of the years spent in the field but also some kind of bias. > What are the manipulations you’re referring to? Need to take a step back and mention that in the field of AI there is a great debate between symbolic and non-symbolic approaches. (and after decades spent with AI under symbolic approaches domination we are now in the golden age of non-symbolic AI; with symbolic starting to have a comeback. this podcast can be a good starting point to learn more https://lexfridman.com/gary-marcus/ https://lexfridman.com/gary-marcus/ - although I disagree with GM on many things - and this tweet for learning about symbolic making a comeback https://twitter.com/hardmaru/status/1470847417193209856 https://twitter.com/hardmaru/status/1470847417193209856) Basically embeddings are "non-symbolic AI" (which is great and this is where their huge potential stems from), but the very way they are generated and then later utlized is completely "symbolic". Which means the the limits of embeddings is defined by the limits of (in this case human written) symbols used to define them. Hope that makes sense.