So we now have models with 0.5 trillion parameters, each the weight of a connection in a neural network.
Trillion-parameter models are surely within reach in the near term -- and that's only within two orders of magnitude of the number of synapses in the human brain, which is in the hundreds of trillions, give or take. To paraphrase the popular saying, a trillion here, a trillion there, and pretty soon you're talking really big numbers.
I know the figures are not comparable apples-to-apples, but still, I find myself in awe looking at how far we've come in just the last few years, to the point that we're realistically contemplating the possibility of seeing dense neural networks with hundreds of trillions of parameters used for real-world applications in our lifetime.
I don't understand this kind of comment. To my mind what it amounts to is "look at how big it is". Alright. So it's big. So what? Is this an elephant pageant?
Suppose a friend comes over and says "I went for dinner at a restaurant. Oh my god the portions were sooo big!". Wouldn't you want to know more information about the food and the restaurant, before you decided whether you're interested in it?
I appreciate that "big" is in peoples' minds associated with "strong", but most of the work in making language models bigger and bigger goes against the normal trend in computer science [1] and also neural networks research in gneral where the trend is to constantly try to reduce the size of models and improve their data efficiency.
What's worse, the trend to supersize language models is never justified, either theoretically (ha ha) or empirically in the relevant literature - and when rival teams make the obvious experiments the evidence is that size is not required to achieve good performance. For example:
It’s Not Just Size That Matters: Small Language Models Are Also Few-Shot Learners
The biggest lesson that can be read from 70 years of AI research is that general methods that leverage computation are ultimately the most effective, and by a large margin. The ultimate reason for this is Moore's law, or rather its generalization of continued exponentially falling cost per unit of computation. Most AI research has been conducted as if the computation available to the agent were constant (in which case leveraging human knowledge would be one of the only ways to improve performance) but, over a slightly longer time than a typical research project, massively more computation inevitably becomes available. Seeking an improvement that makes a difference in the shorter term, researchers seek to leverage their human knowledge of the domain, but the only thing that matters in the long run is the leveraging of computation. These two need not run counter to each other, but in practice they tend to. Time spent on one is time not spent on the other. There are psychological commitments to investment in one approach or the other. And the human-knowledge approach tends to complicate methods in ways that make them less suited to taking advantage of general methods leveraging computation.
We have to learn the bitter lesson that building in how we think we think does not work in the long run. The bitter lesson is based on the historical observations that 1) AI researchers have often tried to build knowledge into their agents, 2) this always helps in the short term, and is personally satisfying to the researcher, but 3) in the long run it plateaus and even inhibits further progress, and 4) breakthrough progress eventually arrives by an opposing approach based on scaling computation by search and learning.
One thing that should be learned from the bitter lesson is the great power of general purpose methods, of methods that continue to scale with increased computation even as the available computation becomes very great. The two methods that seem to scale arbitrarily in this way are search and learning.
A key related question -- to which no one has the answer today -- is whether we must scale computation to match or exceed that of the human brain to be able to replicate or surpass its cognitive abilities. (Note that this question is independent of whether doing so would require future theoretical breakthroughs -- another question to which no one knows the answer today.)
Yes, I've read "The Bitter Lesson". Have you read "A better lesson", by Rodney Brooks?
Edit:
> A key related question -- to which no one has the answer today -- is whether we must scale computation to match or exceed that of the human brain to be able to replicate or surpass its cognitive abilities.
What "computation" is that? Are you talking about scaling up neural networks, which is more in the context of the conversation, but requires some very big assumptions about (artificial) neural networks? Do you mean a different kind of computation?
(Note: my comment, plus the above edit, is a series of questions and I recognise that commens like that can come across as standoffish. This is not my intention, so please accept the questions above as having been asked in the most neutral tone as possible and in the interest of promoting conversation, rather than confrontation.)
Are you talking about scaling up neural networks, which is more in the context of the conversation, but requires some very big assumptions about (artificial) neural networks?
Yes. But note that under the rubric of "deep neural networks" or "deep learning," I would include a lot of things, including combinations of methods like "deep reinforcement learning," learning by self-play via gradual evolution of surviving models, models that use "dense associative memories," of which transformers are only one special case, and future deep learning methods that have not yet been discovered.
And yes, some very big assumptions are required!
FWIW, your comments did not come across as standoffish to me :-)
Except that every synapse is not a dumb weight but a highly complex system connected to an even more complex system (aka neuron) which might each be a (super)computer on its own.
Given how extremely bad we are at computing, there is hope (for ai) that the neurons or their circuits are not _that_ powerful after all.
Emulating a neuron != taking a comparable part in a computation. Probably the former is a lot more complex. For instance, an artificial net can take advantage of backpropagation in a separated training phase -- that's a lot of complexity that's factored out of the runtime phase.
Wonder how this architecture is limiting the space, though - all biological brains train continuously. Our DNNs are more like a brain upload snapshot that's always run for one cycle and then rebooted.
Yeah, that's worth exploring more. It's just that it's not safe to depend on "real neurons are complicated, and therefore artificial nets of simple units won't have transformative capabilities".
Last I heard (and I believe this could be wrong) my professor said that we basically understand how a single neuron works. That like basically if we do X input we get Y output, up to some accuracy. He used this to discuss the idea behind neural networks -- that each neuron is simple enough to model, all we need to worry about is the weights and the dynamics of the network as a whole.
How much of a simplification is that? And how much does the accuracy of such a model matter, in the grand scheme of things?
oh finally something that I learned a lot about :)
Such research is the area of computational neuroscience - one thing that such people do is try to model parts of the brain (or just a single neuron) with computers.
A Neuron (=nerve cell in the brain) is a very complex beast. In rough terms they work like this: They collect signals (electrical impulses) via their small appendages called dendrites. when the sum of the signals reaches a certain threshold a large electrical impulse is generated at the cell body that will travel trough its "output" appendage (called axon) that connected to another neuron's cell body or to its dendrite.
Neurons display a dazzling variety in all these parameters:
- In morphology, e.g. they can look like a pine tree http://www.scholarpedia.org/article/Pyramidal_neuron (I really recommend scholarpedia, also this article has a nice animation on how electrical impulses propagate) or like a sea urchin.
- it really matters where the cell gets its impulse from: A neuron stimulated near its cell body will be much more sensitive to the input than being stimulated far away.
- Their response characteristics are wildly varied too. Some give off one large impulse, some a quick burst of impulses. Some are preventing others from giving out impulses from stimulation (inhibitor neurons)
- This whole mess can be modulated with chemical compounds that are released by the body -- some make some neurons more sensitive, some less.
- Also we still discover every year some new mechanism that modulates how they function.
The issue is that this results in such a complex system that a modern PC cant even simulate 1 detailed neuron model realtime (these tools are open source, try them out! for example https://neuron.yale.edu/ ). Now we know that we're simulating things that likely do not matter (e.g. we don't need a neuron model that consist of 10.000+ segments), but we do not know which parts we need to remove to have a faithful simulation. Also we might simply simulate some parts wrong because our knowledge of the subject is not enough.
But on the upside we've reached some great things already, for example we know how our brain calculates from our head and eye position the orientation of the things we're looking at
All we need to worry about is the weights and the dynamics of the network as a whole. How much of a simplification is that?
A lot. Parallel optimization is an art form. These models are trained on static datasets, they can't intervene in the environment to infer causal relations, so they need legs and hands.
I would say quite a bit. Adding even a third body makes it impossible to calculate physics with certainty. A complex system with any number of individual components is hard to understand with certainty and/or calculations can become exponentially more complex .
The intent was to illustrate that complexities of a system of simple components can be pretty difficult. Automata theory has more appropriate examples perhaps.
That’s correct understanding as of 1943 when the “artificial neural network“ model your professor is teaching was developed.
There is a whole lot of new knowledge on how live neurons and networks of neurons work that had been collected in the last 75 years in the neuroscience domain but it’s mostly ignored by computer scientists.
My issue with this kind of reasoning is the comparison and reference to the human brain. The potential and reach of AI transcends the brain. We never had to master the "mystery" of how birds fly to invent aviation. It was never necessary to compare the number of turbine revolutions of early airplanes to the number of an eagle's feathers. Maybe birds were an inspiration or a metaphor, but thankfully aviation has not been limited to the means of propulsion of the beautiful yet humble pigeon. The potential of aviation has taken us into space exploration and massive international travel. I don't know where AI will take us, but I don't think it will be constrained by this temporary organ called 'human brain'.
If retrieval based NLP [0] becomes a thing, then trillion plus parameters models will likely be less of a thing; as very likely, most of these tens to hundreds of billions of parameters are likely over-fitting (better word: memorized) on training data [text corpus] as seen in the case of GPT-3.
Yes, self-attention mechanisms are dense associative memories, so it might be possible to replace them in many cases with simpler storage mechanisms. Still, I would count the required storage space as part of a model's parameter size -- e.g., a model consisting of 1 trillion values in RAM and 99 trillion values in storage consists of... 100 trillion values.
There was a result recently of modeling an organic neuron with 1000 digital neurons.
And even if that result was perfect modeling of the neuron, that assumes perfect and exhaustive data readings on the organic neuron, which is, frankly, unlikely. (Not that I know how to estimate how much it's missing, but I don't think we fully understand a single neuron yet.)
Expensive is a really relative term right about now... 64gb DDR4 LRDIMMs could be had for about 250$ each on eBay before the chip shortage. While that price is a "good-ish deal", it really wasn't unheard of. A search I just did returns more than a few hits...
Comments
So we now have models with 0.5 trillion parameters, each the weight of a connection in a neural network.
Trillion-parameter models are surely within reach in the near term -- and that's only within two orders of magnitude of the number of synapses in the human brain, which is in the hundreds of trillions, give or take. To paraphrase the popular saying, a trillion here, a trillion there, and pretty soon you're talking really big numbers.
I know the figures are not comparable apples-to-apples, but still, I find myself in awe looking at how far we've come in just the last few years, to the point that we're realistically contemplating the possibility of seeing dense neural networks with hundreds of trillions of parameters used for real-world applications in our lifetime.
We sure live in interesting times.
I don't understand this kind of comment. To my mind what it amounts to is "look at how big it is". Alright. So it's big. So what? Is this an elephant pageant?
Suppose a friend comes over and says "I went for dinner at a restaurant. Oh my god the portions were sooo big!". Wouldn't you want to know more information about the food and the restaurant, before you decided whether you're interested in it?
I appreciate that "big" is in peoples' minds associated with "strong", but most of the work in making language models bigger and bigger goes against the normal trend in computer science [1] and also neural networks research in gneral where the trend is to constantly try to reduce the size of models and improve their data efficiency.
What's worse, the trend to supersize language models is never justified, either theoretically (ha ha) or empirically in the relevant literature - and when rival teams make the obvious experiments the evidence is that size is not required to achieve good performance. For example:
It’s Not Just Size That Matters: Small Language Models Are Also Few-Shot Learners
https://aclanthology.org/2021.naacl-main.185/
____________
[1] Imagine someone bragging that their mergesort implementation has a million LOC! People brag about implementations in few lines of code, not many.
What? That's absurd. Large language models are motivated by empirical scaling law. It is actually better justified than other ML research.
Scaling Laws for Neural Language Models: https://arxiv.org/abs/2001.08361
In my mind, this is now called Pakled reasoning, from the scene in Star Trek: Lower Decks.
https://www.youtube.com/watch?v=lv1uhAa_M_U&t=193sIt's a good question. A while ago, Rich Sutton wrote a good answer for it.: http://incompleteideas.net/IncIdeas/BitterLesson.html -- I recommend reading the whole essay. Quoting him (emphasis mine):
A key related question -- to which no one has the answer today -- is whether we must scale computation to match or exceed that of the human brain to be able to replicate or surpass its cognitive abilities. (Note that this question is independent of whether doing so would require future theoretical breakthroughs -- another question to which no one knows the answer today.)
PS. See also sanxiyn's response: https://news.ycombinator.com/item?id=28838745
Yes, I've read "The Bitter Lesson". Have you read "A better lesson", by Rodney Brooks?
Edit:
What "computation" is that? Are you talking about scaling up neural networks, which is more in the context of the conversation, but requires some very big assumptions about (artificial) neural networks? Do you mean a different kind of computation?
(Note: my comment, plus the above edit, is a series of questions and I recognise that commens like that can come across as standoffish. This is not my intention, so please accept the questions above as having been asked in the most neutral tone as possible and in the interest of promoting conversation, rather than confrontation.)
Yes. But note that under the rubric of "deep neural networks" or "deep learning," I would include a lot of things, including combinations of methods like "deep reinforcement learning," learning by self-play via gradual evolution of surviving models, models that use "dense associative memories," of which transformers are only one special case, and future deep learning methods that have not yet been discovered.
And yes, some very big assumptions are required!
FWIW, your comments did not come across as standoffish to me :-)
New AI fallacy: appeal to size
Except that every synapse is not a dumb weight but a highly complex system connected to an even more complex system (aka neuron) which might each be a (super)computer on its own.
Given how extremely bad we are at computing, there is hope (for ai) that the neurons or their circuits are not _that_ powerful after all.
Emulating a neuron != taking a comparable part in a computation. Probably the former is a lot more complex. For instance, an artificial net can take advantage of backpropagation in a separated training phase -- that's a lot of complexity that's factored out of the runtime phase.
Wonder how this architecture is limiting the space, though - all biological brains train continuously. Our DNNs are more like a brain upload snapshot that's always run for one cycle and then rebooted.
Yeah, that's worth exploring more. It's just that it's not safe to depend on "real neurons are complicated, and therefore artificial nets of simple units won't have transformative capabilities".
Last I heard (and I believe this could be wrong) my professor said that we basically understand how a single neuron works. That like basically if we do X input we get Y output, up to some accuracy. He used this to discuss the idea behind neural networks -- that each neuron is simple enough to model, all we need to worry about is the weights and the dynamics of the network as a whole.
How much of a simplification is that? And how much does the accuracy of such a model matter, in the grand scheme of things?
oh finally something that I learned a lot about :)
Such research is the area of computational neuroscience - one thing that such people do is try to model parts of the brain (or just a single neuron) with computers.
A Neuron (=nerve cell in the brain) is a very complex beast. In rough terms they work like this: They collect signals (electrical impulses) via their small appendages called dendrites. when the sum of the signals reaches a certain threshold a large electrical impulse is generated at the cell body that will travel trough its "output" appendage (called axon) that connected to another neuron's cell body or to its dendrite.
Neurons display a dazzling variety in all these parameters:
- In morphology, e.g. they can look like a pine tree http://www.scholarpedia.org/article/Pyramidal_neuron (I really recommend scholarpedia, also this article has a nice animation on how electrical impulses propagate) or like a sea urchin.
- it really matters where the cell gets its impulse from: A neuron stimulated near its cell body will be much more sensitive to the input than being stimulated far away.
- Their response characteristics are wildly varied too. Some give off one large impulse, some a quick burst of impulses. Some are preventing others from giving out impulses from stimulation (inhibitor neurons)
- This whole mess can be modulated with chemical compounds that are released by the body -- some make some neurons more sensitive, some less.
- Also we still discover every year some new mechanism that modulates how they function.
The issue is that this results in such a complex system that a modern PC cant even simulate 1 detailed neuron model realtime (these tools are open source, try them out! for example https://neuron.yale.edu/ ). Now we know that we're simulating things that likely do not matter (e.g. we don't need a neuron model that consist of 10.000+ segments), but we do not know which parts we need to remove to have a faithful simulation. Also we might simply simulate some parts wrong because our knowledge of the subject is not enough.
But on the upside we've reached some great things already, for example we know how our brain calculates from our head and eye position the orientation of the things we're looking at
A lot. Parallel optimization is an art form. These models are trained on static datasets, they can't intervene in the environment to infer causal relations, so they need legs and hands.
They totally can https://deepmind.com/blog/article/generally-capable-agents-e...
I would say quite a bit. Adding even a third body makes it impossible to calculate physics with certainty. A complex system with any number of individual components is hard to understand with certainty and/or calculations can become exponentially more complex .
Citing the fact that the 3 body problem doesn't always have an exact solution is a straw man argument.
The intent was to illustrate that complexities of a system of simple components can be pretty difficult. Automata theory has more appropriate examples perhaps.
Your professor lied.
That’s correct understanding as of 1943 when the “artificial neural network“ model your professor is teaching was developed.
There is a whole lot of new knowledge on how live neurons and networks of neurons work that had been collected in the last 75 years in the neuroscience domain but it’s mostly ignored by computer scientists.
I’d be really interested in learning more about this. Can you point me to some easily grokable literature?
https://www.amazon.com/Brain-Computations-Edmund-T-Rolls/dp/...
Won't somebody think of the exosomes and telocytes?
My issue with this kind of reasoning is the comparison and reference to the human brain. The potential and reach of AI transcends the brain. We never had to master the "mystery" of how birds fly to invent aviation. It was never necessary to compare the number of turbine revolutions of early airplanes to the number of an eagle's feathers. Maybe birds were an inspiration or a metaphor, but thankfully aviation has not been limited to the means of propulsion of the beautiful yet humble pigeon. The potential of aviation has taken us into space exploration and massive international travel. I don't know where AI will take us, but I don't think it will be constrained by this temporary organ called 'human brain'.
Not all weights are born equal, different paradigms allow more parameters while being less parameter-efficient, e.g. https://openreview.net/forum?id=TXqemS7XEH
If retrieval based NLP [0] becomes a thing, then trillion plus parameters models will likely be less of a thing; as very likely, most of these tens to hundreds of billions of parameters are likely over-fitting (better word: memorized) on training data [text corpus] as seen in the case of GPT-3.
[0] https://ai.stanford.edu/blog/retrieval-based-NLP/
Yes, self-attention mechanisms are dense associative memories, so it might be possible to replace them in many cases with simpler storage mechanisms. Still, I would count the required storage space as part of a model's parameter size -- e.g., a model consisting of 1 trillion values in RAM and 99 trillion values in storage consists of... 100 trillion values.
The switch transformer has already achieved a trillion parameters.
https://arxiv.org/abs/2101.03961
Unless we have misunderstood neurons, and microtubules are the fundamental computational unit in which case we are out by an order of magnitude
There was a result recently of modeling an organic neuron with 1000 digital neurons.
And even if that result was perfect modeling of the neuron, that assumes perfect and exhaustive data readings on the organic neuron, which is, frankly, unlikely. (Not that I know how to estimate how much it's missing, but I don't think we fully understand a single neuron yet.)
CPU in kilohertz then megahertz then gigahertz then it stopped.
RAM in kilobytes then megabytes then gigabytes then it stopped.
Yes for CPU, no for RAM. You can buy a computer with terabytes of RAM just fine. It's just expensive.
Expensive is a really relative term right about now... 64gb DDR4 LRDIMMs could be had for about 250$ each on eBay before the chip shortage. While that price is a "good-ish deal", it really wasn't unheard of. A search I just did returns more than a few hits...
Now compare to single GPU prices...
Those are material science and physical limitations.
Number of parameters in a neural network is not really limited that way, doing useful compute with it is a different matter
China's Wu Dao 2.0 has 1.75 trillion parameters. https://towardsdatascience.com/gpt-3-scared-you-meet-wu-dao-...
A 10 trillion parameter model was mentioned here: https://mobile.twitter.com/ethancaballero/status/14458268620...
That's MoE.
Mixture of Experts, aka not all 10 trillion parameters are used at the same time, just a subset that is an "expert" on the "task at hand".
That would also describe an organic brain though.