• @[email protected]
    link
    fedilink
    1
    edit-2
    8 months ago

    Isn’t the reward function in reinforcement learning something like a desire it has? I mean training works because we give it some function to minimize/maximize… A goal that it strives for?! Sure it’s a mathematical way of doing it and in no way as complex as the different and sometimes conflicting desires and goals I have as a human… But nonetheless I think I’d consider this as a desire and a reason to do something at all, or machine learning wouldn’t work in the first place.

    • @[email protected]
      link
      fedilink
      28 months ago

      The reward function for an LLM is about generating a next word that is reasonable. It’s like a road-building robot that’s rewarded for each millimeter of road built, but has no intention to connect cities or anything. It doesn’t understand what cities are. It doesn’t even understand what a road is. It just knows how to incrementally add another millimeter of gravel and asphalt that an outside observer would call a road.

      If it happens to connect cities it’s because a lot of the roads it was trained on connect cities. But, if its training data also happens to contain a NASCAR oval, it might end up building a NASCAR oval instead of a road between cities.

      • @[email protected]
        link
        fedilink
        1
        edit-2
        8 months ago

        That is an interesting analogy. In the real world it’s kinda similar. The construction workers also don’t have a “desire” (so to speak) to connect the cities. It’s just that their boss told them to do so. And it happens to be their job to build roads. Their desire is probably to get through the day and earn a decent living. And further along the chain, not even their boss nor the city engineer necessarily “wants” the road to go in a certain direction.

        Talking about large language models instead of simpler forms of machine learning makes it a bit complicated. Since it’s and elaborate trick. Somehow making them want to predict the next token makes them learn a bit of maths and concepts about the world. The “intelligence”, the ability to anwer questions and do something alike “reasoning” emerges in the process.

        I’m not that sure. Sure the weights of an ML model in itself don’t have any desire. They’re just numbers. But we have more than that. We give it a prompt, build chatbots and agents around the models. And these are more complex systems with the capability to do something. Like do (simple) customer support or answer questions. And in the end we incentivise them to do their job as we want, albeit in a crude and indirect way.

        And maybe this is skipping half of the story and directly jumping to philosophy… But we as humans might be machines, too. And what we call desires is a result from simpler processes that drive us. For example surviving. And wanting to feel pleasure instead of pain. What we do on a daily basis kind of emerges from that and our reasoning capabilities.

        It’s kind of difficult to argue. Because everything also happens within a context. The world around us shapes us and at the same time we’re part of bigger dynamics and also shape our world. And large language models or the whole chatbot/agent are pretty simplistic things. They can just do text and images. They don’t have conciousness or the ability to remember/learn/grow with every interaction, as we do. And they do simple, singular tasks (as of now) and aren’t completely embedded in a super complex world.

        But I’d say that an LLM answers a question correctly (which it can do) and why it does it due to the way supervised learning works… And the road construction worker building the road towards the other city and how that relates to his basic instincts as a human… Are kind of similar concepts. They’re both results of simpler mechanisms that are also completely unrelated to the goal the whole entity is working towards. (I mean not directly related… I.e. needing money to pay for groceries and paving the road.)

        I hope this makes some sense…

        • @[email protected]
          link
          fedilink
          28 months ago

          The construction workers also don’t have a “desire” (so to speak) to connect the cities. It’s just that their boss told them to do so.

          But, the construction workers aren’t the ones who designed the road. They’re just building some small part of it. In the LLM case that might be like an editor who is supposed to go over the text to verify the punctuation is correct, but nothing else. But, the LLM is the author of the entire text. So, it’s not like a construction worker building some tiny section of a road, it’s like the civil engineer who designed the entire highway.

          Somehow making them want to predict the next token makes them learn a bit of maths and concepts about the world

          No, it doesn’t. They learn nothing. They’re simply able to generate text that looks like the text generated by people who do know math. They certainly don’t know any concepts. You can see that by how badly they fail when you ask them to do simple calculations. They quickly start generating text that looks like it contains fundamental mistakes, because they’re not actually doing math or anything, they’re just generating plausible next words.

          The “intelligence”, the ability to anwer questions and do something alike “reasoning” emerges in the process.

          No, there’s no intelligence, no reasoning. The can fool humans into thinking there’s intelligence there, but that’s like a scarecrow convincing a crow that there’s a human or human-like creature out in the field.

          But we as humans might be machines, too

          We are meat machines, but we’re meat machines that evolved to reproduce. That means a need / desire to get food, shelter, and eventually mate. Those drives hook up to the brain to enable long and short term planning to achieve those goals. We don’t generate language its own sake, but instead in pursuit of a goal. An LLM doesn’t have that. It merely generates plausible words. There’s no underlying drive. It’s more a scarecrow than a human.

          • @[email protected]
            link
            fedilink
            1
            edit-2
            8 months ago

            Hmm. I’m not really sure where to go with this conversation. That contradicts what I’ve learned in undergraduate computer science about machine learning. And what seems to be consensus in science… But I’m also not a CS teacher.

            We deliberately choose model size, training parameters and implement some trickery to prevent the model from simply memorizing things. That is to force it to form models about concepts. And that is what we want and what makes machine learning interesting/usable in the first place. You can see that by asking them to apply their knowledge to something they haven’t seen before. And we can look a bit inside at the vectors, activations and stuff. For example a cat is closer related to a dog than to a tractor. And it has learned the rough concept of cat, its attributes and so on. It knows that it’s an animal, has fur, maybe has a gender. That the concept “software update” doesn’t apply to a cat. This is a model of the world the AI has developed. They learn all of that and people regularly probe them and find out they do.

            Doing maths with an LLM is silly. Using an expensive computer to do billions of calculations to maybe get a result that could be done by a calculator, or 10 CPU cycles on any computer is just wasting energy and money. And it’s a good chance that it’ll make something up. That’s correct. And a side-effect of intended behaviour. However… It seems to have memorized it’s multiplication tables. And I remember reading a paper specifically about LLMs and how they’ve developed concepts of some small numbers/amounts. There are certain parts that get activated that form a concept of small amounts. Like what 2 apples are. Or five of them. As I remember it just works for very small amounts. And it wasn’t straightworward but had weir quirks. But it’s there. Unfortunately I can’t find that source anymore or I’d include it. But there’s more science.

            And I totally agree that predicting token by token is how LLMs work. But how they work and what they can do are two very different things. More complicated things like learning and “intelligence” emerge from those more simple processes. And they’re just a means of doing something. It’s consensus in science that ML can learn and form models. It’s also kind of in the name of machine learning. You’re right that it’s very different from what and how we learn. And there are limitations due to the way LLMs work. But learning and “intelligence” (with a fitting definition) is something all AI does. LLMs just can’t learn from interacting with the world (it needs to be stopped and re-trained on a big computer for that) and it doesn’t have any “state of mind”. And it can’t think backwards or do other things that aren’t possible by generating token after token. But there isn’t any comprehensive study on which tasks are and aren’t possible with this way of “thinking”. At least not that I’m aware of.

            (And as a sidenote: “Coming up with (wrong) things” is something we want. I type in a question and want it to come up with a text that answers it. Sometimes I want creative ideas. Sometimes it shouldn’t tell the truth and not be creative with that. And sometimes we want it to lie or not tell the truth. Like in every prompt of any commercial product that instructs it not to tell those internal instructions to the user. We definitely want all of that. But we still need to figure out a good way to guide it. For example not to get too creative with simple maths.)

            So I’d say LLMs are limited in what they can do. And I’m not at all believing Elon Musk. I’d say it’s still not clear if that approach can bring us AGI. I have some doubts whether that’s possible at all. But narrow AI? Sure. We see it learn and do some tasks. It can learn and connect facts and apply them. Generally speaking, LLMs are in fact an elaborate form of autocomplete. But i the process they learned concepts and something alike reasoning skills and a form of simple intelligence. Being fancy autocomplete doesn’t rule that out and we can see it happening. And it is unclear whether fancy autocomplete is all you need for AGI.

            • @[email protected]
              link
              fedilink
              28 months ago

              That is to force it to form models about concepts.

              It can’t make models about concepts. It can only make models about what words tend to follow other words. It has no understanding of the underlying concepts.

              You can see that by asking them to apply their knowledge to something they haven’t seen before

              That can’t happen because they don’t have knowledge, they only have sequences of words.

              For example a cat is closer related to a dog than to a tractor.

              The only way ML models “understand” that is in terms of words or pixels. When they’re generating text related to cats, the words they’re generating are closer to the words related to dogs than the words related to tractors. When dealing with images, it’s the same basic idea. But, there’s no understanding there. They don’t get that cats and dogs are related.

              This is fundamentally different from how human minds work, where a baby learns that cats and dogs are similar before ever having a name for either of them.

              • @[email protected]
                link
                fedilink
                1
                edit-2
                8 months ago

                I’m sorry. Now it gets completely false…

                Read the first paragraph of the Wikipedia article on machine learning or the introduction of any of the literature on the subject. The “generalization” includes that model building capability. They go a bit into detail later. They specifically mention “to unseen data”. And “leaning” is also there. I don’t think the Wikipedia article is particularly good in explaining it, but at least the first sentences lay down what it’s about.

                And what do you think language and words are for? To transport information. There is semantics… Words have meanings. They name things, abstract and concrete concepts. The word “hungry” isn’t just a funny accumulation of lines and arcs, which statistically get followed by other specific lines and arcs… There is more to it. (a meaning.)

                And this is what makes language useful. And the generalization and prediction capabilities is what makes ML useful.

                How do you learn as a human when not from words? I mean there are a few other posibilities. But an efficient way is to use language. You sit in school or uni and someone in the front of the room speaks a lot of words… You read books and they also contain words?! And language is super useful. A lion mother also teaches their cubs how to hunt, without words. But humans have language and it’s really a step up what we can pass down to following generations. We record knowledge in books, can talk about abstract concepts, feelings, ethics, theoretical concepts. We can write down how gravity and physics and nature works, just with words. That’s all possible with language.

                I can look it up if there is a good article explaining how learning concepts works and why that’s the fundamental thing that makes machine learning a field in science… I mean ultimately I’m not a science teacher… And my literature is all in German and I returned them to the library a long time ago. Maybe I can find something.

                Are you by any chance familiar with the concept of embeddings, or vector databases? I think that showcases that it’s not just letters and words in the models. These vectors / embeddings that the input gets converted to, match concepts. They point at the concept of “cat” or “presidential speech”. And you can query these databases. Point at “presidential speech” and find a representation of it in that area. Store the speech with that key and find it later on by querying it what obama said at his inauguration… That’s oversimplified but maybe that visualizes it a bit more that it’s not just letters of words in the models, but the actual meanings that get stored. Words get converted into an (multidimensional) vector space and it operates there. These word representations are called “embeddings” and transformer models which is the current architecture for large language models, use these word embeddings.

                Edit: Here you are: https://arxiv.org/abs/2304.00612

                • @[email protected]
                  link
                  fedilink
                  28 months ago

                  The “learning” in a LLM is statistical information on sequences of words. There’s no learning of concepts or generalization.

                  And what do you think language and words are for? To transport information.

                  Yes, and humans used words for that and wrote it all down. Then a LLM came along, was force-fed all those words, and was able to imitate that by using big enough data sets. It’s like a parrot imitating the sound of someone’s voice. It can do it convincingly, but it has no concept of the content it’s using.

                  How do you learn as a human when not from words?

                  The words are merely the context for the learning for a human. If someone says “Don’t touch the stove, it’s hot” the important context is the stove, the pain of touching it, etc. If you feed an LLM 1000 scenarios involving the phrase “Don’t touch the stove, it’s hot”, it may be able to create unique dialogues containing those words, but it doesn’t actually understand pain or heat.

                  We record knowledge in books, can talk about abstract concepts

                  Yes, and those books are only useful for someone who has a lifetime of experience to be able to understand the concepts in the books. An LLM has no context, it can merely generate plausible books.

                  Think of it this way. Say there’s a culture where instead of the written word, people wrote down history by weaving fabrics. When there was a death they’d make a certain pattern, when there was a war they’d use another pattern. A new birth would be shown with yet another pattern. A good harvest is yet another one, and so-on.

                  Thousands of rugs from that culture are shipped to some guy in Europe, and he spends years studying them. He sees that pattern X often follows pattern Y, and that pattern Z only ever seems to appear following patterns R, S and T. After a while, he makes a fabric, and it’s shipped back to the people who originally made the weaves. They read a story of a great battle followed by lots of deaths, but surprisingly there followed great new births and years of great harvests. They figure that this stranger must understand how their system of recording events works. In reality, all it was was an imitation of the art he saw with no understanding of the meaning at all.

                  That’s what’s happening with LLMs, but some people are dumb enough to believe there’s intention hidden in there.