3 ms·

Unfortunately, your correction is still misleading because it fails to capture what really sets GPT-3 apart. Ironically, GPT-3's limitation also highlights its

Unfortunately, your correction is still misleading because it fails to capture what really sets GPT-3 apart. Ironically, GPT-3's limitation also highlights its strength. As it is not capable of learning (few shot or otherwise) in the strict sense of permanently changing its parameters based on examples, all its demonstrated capabilities are completely at inference time. It somehow configures itself at inference time so that state machines which produce plausible continuations of whatever pattern it was fed, are most probably generated. This means that whenever it succeeds, it is much more flexible in how it produces its responses. It generalizes on and continues those implicit patterns in the provided input. This paper however, is not replicating that flexibility. Their proposed model is much closer to expectation maximization than it is an instance of what GPT3 does. The novelty of their work, what makes it genuinely useful, is they provide a practical and fairly general way to leverage pre-trained language models to produce classifiers for specific tasks using a very small amount of labeled data. Requiring less effort compared to what would go into fine-tuning. This approach to distillation is an instance of https://en.wikipedia.org/wiki/Semi-supervised_learning https://en.wikipedia.org/wiki/Semi-supervised_learning. Compared to GPT-3, this approach remains at a severe disadvantage when amount of effort and time required to gather data and train something useful is accounted for. On the other hand, if you can fit your problem into the format proposed by the paper, you will likely have more control on the final model's behavior using a small amount of labeled examples (for your specific task), at a significantly lower cost of computation at inference time. Focusing on parameters however, is not even wrong.

leereeves 6y ago
Number of parameters remains an important practical concern, and an important goal for research (see [1] for a recent example). It is true that this model requires a large amount of unlabeled data, in addition to a small amount of labeled data, but gathering unlabeled data is often easy. So I think it's potentially misleading to say "this approach remains at a severe disadvantage when amount of effort and time required to gather data and train something useful is accounted for". I'd say this approach actually has a serious advantage compared to GPT3, which is locked inside the walls of OpenAI, can only be used with their permission, and is too big for most people to use (let alone train) anyway. The cost and effort to use this approach on large real world problems is probably less than using GPT3. And it may also have an advantage in terms of the total amount of data and training required, when all of the data and training in the original pretrained models is included. That is, it may be a more efficient way to train new models from scratch for tasks for which few labeled examples are available. 1: https://ai.googleblog.com/2020/09/advancing-nlp-with-efficient-projection.html https://ai.googleblog.com/2020/09/advancing-nlp-with-efficie...
- Cybiote 6y ago
  Let me preface this by saying I generally agree with what you're saying. My focus below is on the delta. > It is true that this model requires a large amount of unlabeled data, in addition to a small amount of labeled data, but gathering unlabeled data is often easy. Gathering unlabeled data is easier than labeled but can still be challenging. You'll often require careful thought in assembling a distribution of examples. Being able to skip that step yields a significant savings even if not as much as that gained from going from labeled to largely unlabeled data. > I'd say this approach actually has a serious advantage compared to GPT3, which is locked inside the walls of OpenAI, can only be used with their permission, and is too big for most people to use (let alone train) anyway. I fully agree and said as much too. > The cost and effort to use this approach on large real world problems is probably less than using GPT3. I'd say it depends. Most of the effort with GPT3 will involve edge cases. Having a system in front to handle these might eat into labor savings but you could still end up net positive. It's difficult to say without real world data, you might be correct. > That is, it may be a more efficient way to train new models from scratch for tasks for which few labeled examples are available. You're right in general, I think. But it's still worth pointing out GPT3's advantage. It combines a lot of general capabilities, which together with its generative ability and flexibility to input means the level of expertise required to get something useful will be much lower compared to this semi-supervised learning approach. And there are some capabilities it's displayed, one example of many being discussing, querying, pattern matching on computer code that seem hard to replicate with this method.
whimsicalism 6y ago
> It somehow configures itself at inference time so that state machines which produce plausible continuations of whatever pattern it was fed, are most probably generated. What is this sentence supposed to convey? I'm an NLP practicioner/researcher and this isn't even true - as GPT isn't a "state machine" as the latent space is continuous and not finite. Moreover, there is nothing that makes GPT-3 "not capable of learning." It has had very exciting results from language modeling a zero-shot task at inference time, but there's nothing (besides compute) precluding fine-tuning of it in principle. I agree with the rest of your comment.
- Cybiote 6y ago
  > What is this sentence supposed to convey? There are examples where it is able to recognize and continue patterns in strings which if manually generated, would have required a FSM. In fact, some of the more impressive examples would require a stack of some sort so I thought I was rather underselling its capabilities in that arena. > as GPT isn't a "state machine" as the latent space is continuous and not finite. Technically speaking, that is impossible since these models leverage floating point numbers and are limited in memory to whatever hidden and self-attention layers. Practically speaking, in order to generate strings based on patterns as mentioned prior, there must be abstract states which correspond to states and state changes such that thinking in terms of at least state machines is useful. Studying LMs in terms of automata is not strange, there have been papers which do this for specific trained RNNs (such as https://arxiv.org/abs/1711.09576 https://arxiv.org/abs/1711.09576). I contend GPT-3 is capable, to a certain extent, of generating these dynamically at inference time. As far as I know this way of extracting what LMs are doing hasn't been done for Transformers but you can also frame Transformers in term of RNNs so there's no reason why such methods wouldn't readily apply to them too. > Moreover, there is nothing that makes GPT-3 "not capable of learning." I specifically addressed that: to count as learning, without diluting the utility of the term, it has to be capable of remembering. Without permanent changes to its weights, the use of the term learning stretches the word beyond utility.
  - whimsicalism 6y ago
    > Without permanent changes to its weights, the use of the term learning stretches the word beyond utility. Yes, my claim is that there is nothing that makes its weights incapable of being fine-tuned and thus changed.

leereeves 6y ago

Number of parameters remains an important practical concern, and an important goal for research (see [1] for a recent example). It is true that this model requires a large amount of unlabeled data, in addition to a small amount of labeled data, but gathering unlabeled data is often easy. So I think it's potentially misleading to say "this approach remains at a severe disadvantage when amount of effort and time required to gather data and train something useful is accounted for". I'd say this approach actually has a serious advantage compared to GPT3, which is locked inside the walls of OpenAI, can only be used with their permission, and is too big for most people to use (let alone train) anyway. The cost and effort to use this approach on large real world problems is probably less than using GPT3. And it may also have an advantage in terms of the total amount of data and training required, when all of the data and training in the original pretrained models is included. That is, it may be a more efficient way to train new models from scratch for tasks for which few labeled examples are available. 1: https://ai.googleblog.com/2020/09/advancing-nlp-with-efficient-projection.html https://ai.googleblog.com/2020/09/advancing-nlp-with-efficie...

Cybiote 6y ago
Let me preface this by saying I generally agree with what you're saying. My focus below is on the delta. > It is true that this model requires a large amount of unlabeled data, in addition to a small amount of labeled data, but gathering unlabeled data is often easy. Gathering unlabeled data is easier than labeled but can still be challenging. You'll often require careful thought in assembling a distribution of examples. Being able to skip that step yields a significant savings even if not as much as that gained from going from labeled to largely unlabeled data. > I'd say this approach actually has a serious advantage compared to GPT3, which is locked inside the walls of OpenAI, can only be used with their permission, and is too big for most people to use (let alone train) anyway. I fully agree and said as much too. > The cost and effort to use this approach on large real world problems is probably less than using GPT3. I'd say it depends. Most of the effort with GPT3 will involve edge cases. Having a system in front to handle these might eat into labor savings but you could still end up net positive. It's difficult to say without real world data, you might be correct. > That is, it may be a more efficient way to train new models from scratch for tasks for which few labeled examples are available. You're right in general, I think. But it's still worth pointing out GPT3's advantage. It combines a lot of general capabilities, which together with its generative ability and flexibility to input means the level of expertise required to get something useful will be much lower compared to this semi-supervised learning approach. And there are some capabilities it's displayed, one example of many being discussing, querying, pattern matching on computer code that seem hard to replicate with this method.

whimsicalism 6y ago

> It somehow configures itself at inference time so that state machines which produce plausible continuations of whatever pattern it was fed, are most probably generated. What is this sentence supposed to convey? I'm an NLP practicioner/researcher and this isn't even true - as GPT isn't a "state machine" as the latent space is continuous and not finite. Moreover, there is nothing that makes GPT-3 "not capable of learning." It has had very exciting results from language modeling a zero-shot task at inference time, but there's nothing (besides compute) precluding fine-tuning of it in principle. I agree with the rest of your comment.

Cybiote 6y ago
> What is this sentence supposed to convey? There are examples where it is able to recognize and continue patterns in strings which if manually generated, would have required a FSM. In fact, some of the more impressive examples would require a stack of some sort so I thought I was rather underselling its capabilities in that arena. > as GPT isn't a "state machine" as the latent space is continuous and not finite. Technically speaking, that is impossible since these models leverage floating point numbers and are limited in memory to whatever hidden and self-attention layers. Practically speaking, in order to generate strings based on patterns as mentioned prior, there must be abstract states which correspond to states and state changes such that thinking in terms of at least state machines is useful. Studying LMs in terms of automata is not strange, there have been papers which do this for specific trained RNNs (such as https://arxiv.org/abs/1711.09576 https://arxiv.org/abs/1711.09576). I contend GPT-3 is capable, to a certain extent, of generating these dynamically at inference time. As far as I know this way of extracting what LMs are doing hasn't been done for Transformers but you can also frame Transformers in term of RNNs so there's no reason why such methods wouldn't readily apply to them too. > Moreover, there is nothing that makes GPT-3 "not capable of learning." I specifically addressed that: to count as learning, without diluting the utility of the term, it has to be capable of remembering. Without permanent changes to its weights, the use of the term learning stretches the word beyond utility.
- whimsicalism 6y ago
  > Without permanent changes to its weights, the use of the term learning stretches the word beyond utility. Yes, my claim is that there is nothing that makes its weights incapable of being fine-tuned and thus changed.