3 ms·

I think you might be downplaying the scientific method involved in development of novel ML models and methods. It's not exactly a random walk process. Most of t

by p1esk 6y ago

I think you might be downplaying the scientific method involved in development of novel ML models and methods. It's not exactly a random walk process. Most of the progress has been the result of people trying to model what's going on in our heads: convnets modeling our vision system, rnns modeling feedback loops, reinforcement learning modeling sparse reward signals, attention based models modeling, well, attention. Network training methods (e.g. SGD) are based on optimization theory. There are plenty of theories trying to explain why or how things work in deep learning. Most of these are probably wrong, but some, sooner or later, will turn out to be right. Not unlike physics which for years had competing theories (e.g. string theory vs quantum gravity, etc). The original transformer experiment tested the hypothesis: "if we organize layers of attention mechanisms in a certain way and feed them tons of text they will be able to process that text effectively enough to build very rich and consistent language model". GPT experiments test the hypothesis "if we feed the transformer more data its language model might become rich and consistent enough to produce human level results". To me this sounds like an well defined scientific experiment. Using your analogy, GPT-3 is more like if you devised an algorithm which produces n + k of pi digits after processing n pi digits - without knowing anything about how to compute pi, or what pi is. To me that deserves falling over my chair in amazement.

YeGoblynQueenne 6y ago
>> Most of the progress has been the result of people trying to model what's going on in our heads: convnets modeling our vision system, rnns modeling feedback loops, reinforcement learning modeling sparse reward signals, attention based models modeling, well, attention. The problem with all those advances is that they all happened more than 20 years ago. My comment discusses the state of machine learning research right now, which is that there are very few new ideas and the majority of the field doesn't have a clear direction. Note also that the advances you describe were not "guided by theory". They were inspired by ideas about how the mind works. But, finding ideas to try is not the scientific process I describe above. And just because you're inspired by an idea doesn't mean your work is in any way a proof or disproof of that idea. For example, CNNs were not created in an effort to demonsrate the accuracy of a certain model of the visual cortex. In fact, Yan LeCun is on record saying that deep learning is nothing like the brain: Yann LeCun: My least favorite description [of deep learning] is, “It works just like the brain.” I don’t like people saying this because, while Deep Learning gets an inspiration from biology, it’s very, very far from what the brain actually does. https://spectrum.ieee.org/automaton/artificial-intelligence/machine-learning/facebook-ai-director-yann-lecun-on-deep-learning https://spectrum.ieee.org/automaton/artificial-intelligence/... >> Using your analogy, GPT-3 is more like if you devised an algorithm which produces n + k of pi digits after processing n pi digits - without knowing anything about how to compute pi, or what pi is. That's not a good example. GPT-3 can't actually do this. In fact, no technology we know of can do this for a k sufficiently large and with accuracy better than chance. Personally I find GPT-3's text generation very underwhelming and not anywhere near the magickal guessing machine your example seems to describe.
YeGoblynQueenne 6y ago
Specifically about attention: >> The original transformer experiment tested the hypothesis: "if we organize layers of attention mechanisms in a certain way and feed them tons of text they will be able to process that text effectively enough to build very rich and consistent language model". I read the paper when it came out and I checked it again now to confirm: there's no such hypothesis in there. In fact "Attention is all you need" is a typical example of a post-hoc paper that describes what a research team did that worked and how well it worked. "We tweaked these knobs and out came STUFF!". Typically for deep learning papers it lacks a theory section, the space of which is instead taken by an "Architecture" section wich well, describes the architecture. There are no theorems, or proofs. There is nothing that connects some kind of theoretical claim to the experiments. The main claim of the paper is "we build this system and it has better performance than previous systems". Like I say in my earlier comment, that's not an interesting scientific claim. I'm sorry but personally I find that kind of work irritating. "We tried some stuff and got some results". Woo-hoo. But, why did you try that stuff and why did you get those results? Did you try to get the same results without that stuff? Did you try to get some other results with the same stuff? Can you explain what is going on in your system and why it does that when I twist this knob? If I twist that knob, can you tell me what it will do without having to run it first to find out? Typically, 99% of the time, the answer to all this is "no" (i.e. no ablation experiments, no theoretical explanations, etc. etc, no nothing). It's like I say above, just throwing stuff at the wall to see what sticks. And then writing a paper to describe it. Oh, and calling stuff suggestive names like "attention". If I call it "boredom", am I more or less justified than the authors?

by p1esk 6y ago

YeGoblynQueenne 6y ago

>> Most of the progress has been the result of people trying to model what's going on in our heads: convnets modeling our vision system, rnns modeling feedback loops, reinforcement learning modeling sparse reward signals, attention based models modeling, well, attention. The problem with all those advances is that they all happened more than 20 years ago. My comment discusses the state of machine learning research right now, which is that there are very few new ideas and the majority of the field doesn't have a clear direction. Note also that the advances you describe were not "guided by theory". They were inspired by ideas about how the mind works. But, finding ideas to try is not the scientific process I describe above. And just because you're inspired by an idea doesn't mean your work is in any way a proof or disproof of that idea. For example, CNNs were not created in an effort to demonsrate the accuracy of a certain model of the visual cortex. In fact, Yan LeCun is on record saying that deep learning is nothing like the brain: Yann LeCun: My least favorite description [of deep learning] is, “It works just like the brain.” I don’t like people saying this because, while Deep Learning gets an inspiration from biology, it’s very, very far from what the brain actually does. https://spectrum.ieee.org/automaton/artificial-intelligence/machine-learning/facebook-ai-director-yann-lecun-on-deep-learning https://spectrum.ieee.org/automaton/artificial-intelligence/... >> Using your analogy, GPT-3 is more like if you devised an algorithm which produces n + k of pi digits after processing n pi digits - without knowing anything about how to compute pi, or what pi is. That's not a good example. GPT-3 can't actually do this. In fact, no technology we know of can do this for a k sufficiently large and with accuracy better than chance. Personally I find GPT-3's text generation very underwhelming and not anywhere near the magickal guessing machine your example seems to describe.

YeGoblynQueenne 6y ago

Specifically about attention: >> The original transformer experiment tested the hypothesis: "if we organize layers of attention mechanisms in a certain way and feed them tons of text they will be able to process that text effectively enough to build very rich and consistent language model". I read the paper when it came out and I checked it again now to confirm: there's no such hypothesis in there. In fact "Attention is all you need" is a typical example of a post-hoc paper that describes what a research team did that worked and how well it worked. "We tweaked these knobs and out came STUFF!". Typically for deep learning papers it lacks a theory section, the space of which is instead taken by an "Architecture" section wich well, describes the architecture. There are no theorems, or proofs. There is nothing that connects some kind of theoretical claim to the experiments. The main claim of the paper is "we build this system and it has better performance than previous systems". Like I say in my earlier comment, that's not an interesting scientific claim. I'm sorry but personally I find that kind of work irritating. "We tried some stuff and got some results". Woo-hoo. But, why did you try that stuff and why did you get those results? Did you try to get the same results without that stuff? Did you try to get some other results with the same stuff? Can you explain what is going on in your system and why it does that when I twist this knob? If I twist that knob, can you tell me what it will do without having to run it first to find out? Typically, 99% of the time, the answer to all this is "no" (i.e. no ablation experiments, no theoretical explanations, etc. etc, no nothing). It's like I say above, just throwing stuff at the wall to see what sticks. And then writing a paper to describe it. Oh, and calling stuff suggestive names like "attention". If I call it "boredom", am I more or less justified than the authors?