Live data from Hacker News

Von Neumann’s critique of automata theory and logic in computer science (1947)

yodaiken.com

31–35 of 35 posts

Re: Von Neumann’s critique of automata theory and logic in computer science (1947)

#31
post #29

Earlier quoted context omitted.

> You can't have a theory of computation without mathematics. You're right that you can't have a theory without some language to talk about it (and it also requires people to come up with it, so is psychology more fundamental than physics?), but the theory of computation is about the laws that govern the power of mathematics. I.e. mathematics (and the universe) was constrained by computation long before anyone knew t…

Psychology and physics are two independent and different fields. Physics wasn't built on top of psychology and vice versa. What language of psychology do you need to talk about physics and vice versa? Newtonian physics and einstein physics or quantum physics may be better examples? Or psychology with developmental psychology, behavioral psychology, etc. Are you saying Turing, the world famous mathematician, didn't us…

> Are you saying Turing, the world famous mathematician, didn't use any mathematical ideas?

Of course he did, but his ideas also came to him through psychology and he expressed them in English. That doesn't make psychology or English more fundamental than computation. Because humans create theories and humans are very complex, almost everything human is involved in the construction of theories, but when we talk about something being more fundamental than another we're not talking about the human process of the theory's construction but about the subject matter of the theory. The theory of computation is not only concerned with matters at a "lower-level" than mathematics, but also lower than logic.

In fact, that computation can be described using mathematics (or English) is precisely because of Turing's discovery of universal computation, which means that any system of symbols that's "rich enough" can describe any other.

> You do realize that his 1936 was filled with mathematical proofs?

Ah, but you should take a closer look at them. His proof of what we today know as the halting theorem goes to great lengths to avoid using any logical axioms. There's a great paper about that by the Turing scholar, Juliet Floyd (https://mdetlefsen.nd.edu/assets/201037/jf.turing.pdf esp. §4.5). He did that because he was trying to get to an idea that's even more fundamental than logic.

> Just because it isn't full of numbers doesn't mean it's not mathematics.

As I said, some people do consider formal logic, and even computation as branches of mathematics (though others don't), but if so you can think of computation as more fundamental than any other branch. Let me put this more precisely instead of speaking in the abstract: computation is more fundamental than the natural numbers (the axiom of infinity, often considered the most basic mathematical axiom) and even the most basic axioms of logic, such as the principle of explosion, in the sense that neither can be given a precise sense without computation, but computation is described without them.

Re: Von Neumann’s critique of automata theory and logic in computer science (1947)

#32
post #21

Earlier quoted context omitted.

> I've often wondered what a theory of computation based in differential equations would look like. The theory of computation is not based on anything ; it is fundamental in the sense that it exposes laws of nature (it is heavily based on physical limitations) independent of the means of talking about them. In a very strong sense it is more fundamental than mathematics itself as it is concerned with how hard it is to…

Theory of computation is based in mathematics/logic. It is a part of mathematics just like computer science is a field of mathematics. I'm not sure I agree that it is more fundamental than mathematics since it is a part of mathematics. You can't have a theory of computation without mathematics. Just like you can't have the field of cryptography without mathematics. Theory of computation and cryptography are built on…

From a modern physics perspective, pron is correct in some sense. To a physicist, the theory of computation and complexity, are built on top of the laws of the universe that you live in. You change the laws of the universe and the difficulty of computation changes. In fact, more generally, what information processing tasks (such as cryptography) are possible or their difficulty are determined by the laws of physics. For example, its possible to copy unknown information in the Newtonian universe, but generally impossible in the quantum universe. Or for example, quantum computers are said to have different complexity than classical computers.

In fact, significant progress in physics in recent times has come about because people ask what sort of information processing tasks should be possible/easy in our universe and which ones not. This allows us to reject physical theories that don't respect our intuitions about information processing. What I am getting at is that the theory of information processing (computation included) is not divorced from reality but can be divorced from mathematics. Even if we completely changed our mathematical systems (start from different axioms than ZFC for example), the type of computations possible in our universe would not change. In other words, if we wanted to use the new mathematics to model computers/information processing systems in our universe, that mathematics would have to respect the information processing results we already know about our universe.

Re: Von Neumann’s critique of automata theory and logic in computer science (1947)

#33
post #25
post #16

can someone provide an ELI5 for this critique?

The "normal" physical systems we build (say houses and hydraulic systems) are mostly "continuous". This means small changes in the inputs generally result in small changes in the behaviour of the system. So, for example, if the screws on your kitchen cabinet doors are not tightened quite enough, the cabinet doors might wobble a little -- that is, a small change resulted in a small change. On the other hand, systems b…

Great summary.

Re: Von Neumann’s critique of automata theory and logic in computer science (1947)

#34

Controls engineer here. I've often wondered what a theory of computation based in differential equations would look like. We desperately need one. Control theory gives us some tools, but not nearly powerful enough ones. It's fairly clear by now that AI, at least as applied to domains immersed in the natural world, such as robotics and machine vision, is better thought of as analog in nature. Ten years ago I'd go to r…

> Ten years ago I'd go to robotics conferences and it would not be unusual for me to be the only person in the room who knew that control theory was a thing.

This is mind-blowing to me. When I thought I might have a reason to develop some robotics technology, the very first thing I did was to get a control theory book, on the off chance that I was really going to do this in the future. In fact, I think this explains why the product I think could be made does not exist yet.

Re: Von Neumann’s critique of automata theory and logic in computer science (1947)

#35

Controls engineer here. I've often wondered what a theory of computation based in differential equations would look like. We desperately need one. Control theory gives us some tools, but not nearly powerful enough ones. It's fairly clear by now that AI, at least as applied to domains immersed in the natural world, such as robotics and machine vision, is better thought of as analog in nature. Ten years ago I'd go to r…

If you literally want a visual representation, there are systems of differential equations that model biological/chemical reaction-diffusion. Roughly speaking, this gives you a bridge between cellular automata, reaction-diffusion and morphogenesis by defining chemicals, states, species - whatever - as a system of differential equations. These produce organic and natural looking patterns!
Post reply on HN