I'm pretty sure getting computers to do arithmetic is not a giant hurdle.
Transformers Can Do Arithmetic with the Right Embeddings
181–190 of 215 posts
Re: Transformers Can Do Arithmetic with the Right Embeddings
#182Earlier quoted context omitted.
Isn't it more that they don't have ready access to the much-more-fundamental concept of decimal numbers? My understanding was that they tokenized them into chunks and tried to learn associations between the chunks, the same as if one was breaking apart English words. So "2+2=4" isn't being treated that differently from "all's well that ends well." This might lead to a kind of Benny's Rules [0] situation, where suffic…
Fascinating article!
It looks like the math-notation formatting didn't survive, for that you might want to see a PDF, ex: https://people.wou.edu/~girodm/library/benny.pdf
Re: Transformers Can Do Arithmetic with the Right Embeddings
#183Earlier quoted context omitted.
> Problem is the current systems can’t reason about things Sounds like the AGI argument trap: They're not able to reason, but we can't succintly define what it is. I don't come with a reasoning chip. Whatever I call reasoning happens as a byproduct of my neural process. I do think that the combination of a transformer network and calls to customized reasoning chips (systems that search and deduce answers, like Wolfra…
I suppose it's a question whether what we call "reasoning" is an emergent phenomenon from having enough connections in a graph, or whether it's some other special sauce which we simply don't have in our current models yet. E.g. humans follow a deductive process to answer questions which they haven't encountered yet. Do we gain this ability purely from a denser/larger graph of knowledge, or from a completely different…
All of our deductive reasoning is founded in induction. For example, the basis of all arithmetic is physics analogies regarding things that exist and the understanding that a thing implies another thing is not based in deduction. Similarly, I suspect from my own experience that general reasoning requires a basic understanding of physics if its origin isn't something ineffable. The ability to connect and find implications cannot itself be purely deductive and it would seem to me that an understanding of physical reality would have to be the origin for that ability.
Re: Transformers Can Do Arithmetic with the Right Embeddings
#184Earlier quoted context omitted.
>> deductive reasoning is just drawing specific conclusion from general patterns. This is according to whom, please?
The fundamental argument of "Artificial Intelligence, Natural Stupidity" is that AI researchers constantly abuse terms like "reasoning," "deduction," "understanding," and so on, deluding others and themselves that their machine is almost as intelligent as a human when it's clearly dumber than a dog. My cats don't need "general patterns" to form deductions, they deduce many sophisticated things (on their terms) with n…
Ants behave in ways that a modern computer still can't imitate. I don't think that generalized intelligence is possible but if it is it would need a different starting point than our current computing hardware. Even insects are flexible in ways that computers aren't.
Re: Transformers Can Do Arithmetic with the Right Embeddings
#185Wow, a lot of grumpiness in here. If it's true that adding like 20 or so tokens to encode column location / decimal spot triples math performance in out of band tasks, that's a big deal. It's a simple fix, it improves performance A LOT, and they even indicate it's not just a party trick, in that the LLM can use the information to do better on related tasks like sorting and list making. This is basically free to add,…
It's also obvious and it's hacky. Frankly I'm stunned this hasn't been tried yet. The people thinking this is a stepping stone to More Intelligence are missing the forest for the trees. Deep learning is always and only ever about representing data abstractly . The more abstractions you can make irrelevant (why would you have to learn how to do math when the base-10 perspective on ASCII-digits is already provided for…
I'd rather not wait another billion or so years for computers to evolve themselves
Re: Transformers Can Do Arithmetic with the Right Embeddings
#186Earlier quoted context omitted.
engineering vs. science -> scientist-types find such hacks ugly whereas engineers have to pay bills and get things moving fast.
And when engineers accumulate enough related hacks, scientist-types may discover a pattern and find a proper, general solution. But they wouldn't get there without the pile of hacks that are effectively meta-level empirical evidence.
AI research has mostly progressed when there’s been enough processing power to avoid needing to use the old style of hacks rather than any sort of generalization going on.
AlphaZero vs Stockfish wasn’t some outgrowth of existing methods. They basically throw the old style away and started over.
Object recognition, LLM’s etc all involved throwing what used to be unimaginable levels of data and compute at a problem that “suddenly” worked. Not saying the people at OpenAI aren’t clever, but instead that it wouldn’t have worked in 2000.
Re: Transformers Can Do Arithmetic with the Right Embeddings
#187Earlier quoted context omitted.
The point is don't ask an LLM to do tasks that a calculator can do. Ask if to use the calculator, just like most humans would.
The point is that you shouldn't need to ask. If it's actually a general-purpose system it will do it automatically.
It still needs to be trained on what calculators are, on how to use them, and on when it is appropriate to do so. Again, same as humans.
Re: Transformers Can Do Arithmetic with the Right Embeddings
#188Earlier quoted context omitted.
LLMs are deterministic. We just sample the results, no? Also, no reason AI needs not be deterministic.
> LLMs are deterministic. In theory, yes. In practice, parallelism combined with floating point math make current implementations fundamentally non-deterministic.
Not fundamentally, but extremely costly to keep track of everything. Still deterministic though.
Re: Transformers Can Do Arithmetic with the Right Embeddings
#189For things like this where we have computationally cheap, well understood, reliable tools available (aka calculator) it seems better to train the model in tool use. I guess perhaps the techniques could be generalized though?
To me this finding shows how transformers don't generalise, since they need specialised embeddings to handle a problem
The fact that transformers generalize is kinda evident from the fact that they can solve novel puzzles.
Re: Transformers Can Do Arithmetic with the Right Embeddings
#190Earlier quoted context omitted.
> They're not able to reason, but we can't succintly define what it is. For transformer-based LLMs, and most LLMs there's an obvious class of problems that they cannot solve. LLMs generally perform bounded computation per token, so they cannot reason about computational problems that are more than linearly complex, for a sufficiently large input instance. If you have a back-and-forth (many shot) your LLM can possibly…
Humans can realise they don’t understand something and seek more knowledge to learn to understand it. But also humans can build complex structures out of simple fundamentals: The same logic of counting up beans on a table can be extrapolated to multiplying that table of beans. And then counting horses the same way you count beans but give them a value of multiple beans. And then simplify that by trading in promises o…
We’ll have to wait and see how far multi modal training takes us. Text only models are extremely limited by the kind of information we can encode as text and the loss of detail e.g. the word “cat” vs an image of a cat vs video of a cat vs direct physical interaction with a cat vs being a mammal that shares a great deal of biology with a cat. You need a table and beans before you can invent a method for counting them