5 ms·

We read words. To translate some words into other words of different literal meanings while maintaining non-literal metaphorical meaning implies the capacity to

by knome 4y ago

We read words. To translate some words into other words of different literal meanings while maintaining non-literal metaphorical meaning implies the capacity to "understand" the words at that level. That it represents data as tokens is irrelevant to the kinds of processing it obviously performs. Pretending GPT is just using a percent chance that some token should show up next is absurd. It's almost a hundred layers of neural pathways.

dvt 4y ago
> Pretending GPT is just using a percent chance that some token should show up next is absurd. It's almost a hundred layers of neural pathways. That is literally what it's doing though. "Pretending a CPU is just adding 0s and 1s is absurd. It's made out of billions and billions of transistors."
- knome 4y ago
  A multi-layer neural net is not limited to vague statistical combinations. You can represent XOR in a neural network. A NAND. If you can represent NAND. You can represent arbitrary logic. Across billions of weights and across nearly a hundred layers of possible decision points. Meaning can be spread across the neurons in ways that allow each to be part of many many different representations. This is only in GPT-3. GPT-4 and soon 5 no doubt exceed these considerably. https://towardsdatascience.com/how-neural-networks-solve-the-xor-problem-59763136bdd7 https://towardsdatascience.com/how-neural-networks-solve-the...
  - dvt 4y ago
    > A multi-layer neural net is not limited to vague statistical combinations. It, absolutely, unequivocally, 100% is only limited to statistical inference. That's how it works. The problem of linear separability (which is like 60 years old, by the way) has nothing to do with logic, but it's actually a byproduct of (a) the fact that neural networks use linear activation functions, and (b) XOR is not separable by a single hyperplane. Hence, you need multiple layers. No idea how you get to "arbitrary logic" from this.
    - toxik 4y ago
      > neural networks use linear activation functions This is exactly wrong, NNs use nonlinear activation functions. That’s the whole point [in activation functions].
PheonixPharts 4y ago
In the grand scheme of things GPT isn't learning "meaning" in any different sense than Latent Semantic Analysis (LSA) learns meaning (which is just using "percent chance that some token should show up next" using covariance). For those without a background in NLP, LSA is literally SVD/PCA (which is ultimately just based on the covariance matrix) on the term frequency matrix for a corpus of text. However LSA is well known to learn certain meanings such as "pet" by projecting words like "dog" and "cat" onto the same principle component. Semantics is learned by exploring the "loadings" in PCA terminology. Now LSA is just a linear transformation, but the idea is that same as GPT that uses a wildly more sophisticated non-linear transformation: the assumption is that words (or tokens in the case of GPT) are caused by some latent variables (in the case of LSA, for GPT high dimensional embedding) and most importantly the patterns of word/token usage allow us to infer what those latent variables are. So it's not absurd and it's how statistical NLP has always worked, only now we have much, much more sophisticated ways to model that latent space. LSA, HMM, and Transformers all assume that the patterns of text usage can be used to model an unobserved latent space that generates the text, which is what we traditionally call "meaning". Ultimately the limits of GPT-N (unless there are none and it does become AGI) will be valuable in showing the limitations of this fundamental model of semantics.
- knome 4y ago
  When I said it operates on a level of meaning, I refer to its ability to perform transformations like this: H: translate the following sentence to instead use olympic curling related metaphors ( do not explicitly mention curling, and do not use similes in forming the translation ): man:"I've been spinning my tires all day without getting anywhere" woman:"running out of gas?" man:"nah, I'm gassed up, geared up, oiled and revving. I just can't seem to get in the right lane, you know?" woman:"anything knocking?" man:"nah, I'm tuned just fine. I can't seem to get my GPS to connect, you know?" woman:"take a few laps without a destination, see if you can figure out the lay of the land" man:"you know, that might help. thanks barbara" R: man: "I've been sweeping the ice all day without hitting the target." . woman: "Losing momentum?" . man: "Nah, I've got the energy, the technique, and the focus. I just can't seem to find the right path, you know?" . woman: "Anything coming in your way?" . man: "Nah, I'm on point with my aim. I just can't seem to get my strategy to align, you know?" . woman: "Maybe glide a few times without aiming specifically, see if you can figure out the ice conditions." . man: "You know, that might help. Thanks, Barbara." I don't think you can perform this kind of translation by merely comparing tokens statistically tending to be near one another. I understand that this is how the data was encoded and the output judged. However, the neural net between those two points is free to encode whatever interconnecting values it happens to as its training drives it through to whatever multidimensional local maximum it manages to settle upon. In short, I expect that it has to translate to and from an abstract meaning in order to manipulate metaphors in such a manner. Most human language is metaphorical, as discussed in Lakoff & Johnson's "Metaphors We Live By". We have a core of experiential terms, but near the whole of the rest of our language is metaphoric in nature. The better a model gets at "understanding" English, the closer to a metaphor engine it must become, needing to be capable of applying arbitrary metaphors and partial metaphors in order to properly manipulate the language. I'm not saying your position is wrong, but it reminds me of the contrasting viewpoints given in Hofstadter's "I am a Strange Loop" as he discussed his "careenium" and "simm balls" analogy for the rise of complex epiphenomena arising from simple rule bound systems.
jillesvangurp 4y ago
Exactly. Pretending our own brains are much more sophisticated seems increasingly a hard to defend position. Apparently, this is good enough for a lot of stuff we think we are good at already.

by knome 4y ago

dvt 4y ago

> Pretending GPT is just using a percent chance that some token should show up next is absurd. It's almost a hundred layers of neural pathways. That is literally what it's doing though. "Pretending a CPU is just adding 0s and 1s is absurd. It's made out of billions and billions of transistors."

knome 4y ago
A multi-layer neural net is not limited to vague statistical combinations. You can represent XOR in a neural network. A NAND. If you can represent NAND. You can represent arbitrary logic. Across billions of weights and across nearly a hundred layers of possible decision points. Meaning can be spread across the neurons in ways that allow each to be part of many many different representations. This is only in GPT-3. GPT-4 and soon 5 no doubt exceed these considerably. https://towardsdatascience.com/how-neural-networks-solve-the-xor-problem-59763136bdd7 https://towardsdatascience.com/how-neural-networks-solve-the...
- dvt 4y ago
  > A multi-layer neural net is not limited to vague statistical combinations. It, absolutely, unequivocally, 100% is only limited to statistical inference. That's how it works. The problem of linear separability (which is like 60 years old, by the way) has nothing to do with logic, but it's actually a byproduct of (a) the fact that neural networks use linear activation functions, and (b) XOR is not separable by a single hyperplane. Hence, you need multiple layers. No idea how you get to "arbitrary logic" from this.
  - toxik 4y ago
    > neural networks use linear activation functions This is exactly wrong, NNs use nonlinear activation functions. That’s the whole point [in activation functions].

PheonixPharts 4y ago

In the grand scheme of things GPT isn't learning "meaning" in any different sense than Latent Semantic Analysis (LSA) learns meaning (which is just using "percent chance that some token should show up next" using covariance). For those without a background in NLP, LSA is literally SVD/PCA (which is ultimately just based on the covariance matrix) on the term frequency matrix for a corpus of text. However LSA is well known to learn certain meanings such as "pet" by projecting words like "dog" and "cat" onto the same principle component. Semantics is learned by exploring the "loadings" in PCA terminology. Now LSA is just a linear transformation, but the idea is that same as GPT that uses a wildly more sophisticated non-linear transformation: the assumption is that words (or tokens in the case of GPT) are caused by some latent variables (in the case of LSA, for GPT high dimensional embedding) and most importantly the patterns of word/token usage allow us to infer what those latent variables are. So it's not absurd and it's how statistical NLP has always worked, only now we have much, much more sophisticated ways to model that latent space. LSA, HMM, and Transformers all assume that the patterns of text usage can be used to model an unobserved latent space that generates the text, which is what we traditionally call "meaning". Ultimately the limits of GPT-N (unless there are none and it does become AGI) will be valuable in showing the limitations of this fundamental model of semantics.

knome 4y ago
When I said it operates on a level of meaning, I refer to its ability to perform transformations like this: H: translate the following sentence to instead use olympic curling related metaphors ( do not explicitly mention curling, and do not use similes in forming the translation ): man:"I've been spinning my tires all day without getting anywhere" woman:"running out of gas?" man:"nah, I'm gassed up, geared up, oiled and revving. I just can't seem to get in the right lane, you know?" woman:"anything knocking?" man:"nah, I'm tuned just fine. I can't seem to get my GPS to connect, you know?" woman:"take a few laps without a destination, see if you can figure out the lay of the land" man:"you know, that might help. thanks barbara" R: man: "I've been sweeping the ice all day without hitting the target." . woman: "Losing momentum?" . man: "Nah, I've got the energy, the technique, and the focus. I just can't seem to find the right path, you know?" . woman: "Anything coming in your way?" . man: "Nah, I'm on point with my aim. I just can't seem to get my strategy to align, you know?" . woman: "Maybe glide a few times without aiming specifically, see if you can figure out the ice conditions." . man: "You know, that might help. Thanks, Barbara." I don't think you can perform this kind of translation by merely comparing tokens statistically tending to be near one another. I understand that this is how the data was encoded and the output judged. However, the neural net between those two points is free to encode whatever interconnecting values it happens to as its training drives it through to whatever multidimensional local maximum it manages to settle upon. In short, I expect that it has to translate to and from an abstract meaning in order to manipulate metaphors in such a manner. Most human language is metaphorical, as discussed in Lakoff & Johnson's "Metaphors We Live By". We have a core of experiential terms, but near the whole of the rest of our language is metaphoric in nature. The better a model gets at "understanding" English, the closer to a metaphor engine it must become, needing to be capable of applying arbitrary metaphors and partial metaphors in order to properly manipulate the language. I'm not saying your position is wrong, but it reminds me of the contrasting viewpoints given in Hofstadter's "I am a Strange Loop" as he discussed his "careenium" and "simm balls" analogy for the rise of complex epiphenomena arising from simple rule bound systems.

jillesvangurp 4y ago

Exactly. Pretending our own brains are much more sophisticated seems increasingly a hard to defend position. Apparently, this is good enough for a lot of stuff we think we are good at already.