Qudits: The Real Future of Quantum Computing?
spectrum.ieee.org
Qudits: The Real Future of Quantum Computing?
1–10 of 36 posts
Re: Qudits: The Real Future of Quantum Computing?
#2I know these guys understand quantum anything more than I ever will, but is having a larger radix specifically a practical way for increasing performance?
If a binary device has a hardware fault, bits that are in error still have a 50% chance of being correct. It seems like a single component failing here would make this worse (five times worse if it wasn't quantum).
I don't know whether or not the pros outweigh the cons here, or at what radix they do (there's probably some statistical analysis that can answer that), but ten just _seems_ like too much.
Re: Qudits: The Real Future of Quantum Computing?
#3I know these guys understand quantum anything more than I ever will, but is having a larger radix specifically a practical way for increasing performance? If a binary device has a hardware fault, bits that are in error still have a 50% chance of being correct. It seems like a single component failing here would make this worse (five times worse if it wasn't quantum). I don't know whether or not the pros outweigh the…
quantum algorithms change complexity class. The pro is simply too big if it can work in reality.
Re: Qudits: The Real Future of Quantum Computing?
#4You don't actually want qubits, you want an analog computer with differentiable signals. Most likely photonic. Qubits are a dead evolution branch.
I've been recently exploring computational metamaterials for photonic computation.
http://users.ece.utexas.edu/~aalu/research%20-%20page%203.ht... (there's quite a few papers on this but unfortunately they are all paywalled. Spoiler alert, they seem to be based entirely on Fourier transform).
These computational metamaterials don't need electricity to be powered (you need something that will shoot the photons on them and read back the values off tho).
Machine learning would be much, much faster on these as you have O(1) differential calculus.
They don't heat up. You can possibly build a house sized CPU out of these. I can see it, a city block sized CPU and a nuclear reactor next to it.
Did you know that on an analog machine, you can do sort in O(n)?
https://en.wikipedia.org/wiki/Spaghetti_sort
Hit me up if you wanna chat about this. I've seen the "light" (xdddd) now and can't go back to stupid bits.
I'm not like super married to the metamaterials but analog photonic trumps quantum for just about every task I can think of.
Re: Qudits: The Real Future of Quantum Computing?
#5How about.. oh i don't know... actually building a quantum computer?? no?? ok. Sry..
Re: Qudits: The Real Future of Quantum Computing?
#6You don't actually want qubits, you want an analog computer with differentiable signals. Most likely photonic. Qubits are a dead evolution branch. I've been recently exploring computational metamaterials for photonic computation. http://users.ece.utexas.edu/~aalu/research%20-%20page%203.ht... (there's quite a few papers on this but unfortunately they are all paywalled. Spoiler alert, they seem to be based entirely on…
> You don't actually want qubits, you want an analog computer with differentiable signals. Most likely photonic. Qubits are a dead evolution branch.
that's quite a bold claim. im pretty intrigued, but it seems like there isn't a lot of easily accessible information on this topic. the wikipedia article on analog computers mostly focuses on mechanical ones from the past. do you have any recommended starter readings for analog computation? is it a model that has the computational time complexity of a non-deterministic turing machine? thanks.
Re: Qudits: The Real Future of Quantum Computing?
#7You don't actually want qubits, you want an analog computer with differentiable signals. Most likely photonic. Qubits are a dead evolution branch. I've been recently exploring computational metamaterials for photonic computation. http://users.ece.utexas.edu/~aalu/research%20-%20page%203.ht... (there's quite a few papers on this but unfortunately they are all paywalled. Spoiler alert, they seem to be based entirely on…
This may be a rather naive question, as I've never seriously considered a modern take on analog computers with differentiable signals, but intuitively it's easy to see qualitative advantages in algorithms based on differential calculus, Fourier analysis, etc. What's less obvious to me is how such machines would do at fundamentally discrete tasks.
For a trivially simple concrete example, consider the task of computing the results of the SHA-256 hashing algorithm, which is primarily bottlenecked by bit shifts and rotates on integers. How would your analog machine compare to conventional digital computers? Is there any literature on this?
Re: Qudits: The Real Future of Quantum Computing?
#8You don't actually want qubits, you want an analog computer with differentiable signals. Most likely photonic. Qubits are a dead evolution branch. I've been recently exploring computational metamaterials for photonic computation. http://users.ece.utexas.edu/~aalu/research%20-%20page%203.ht... (there's quite a few papers on this but unfortunately they are all paywalled. Spoiler alert, they seem to be based entirely on…
I don't think you understand why quantum computers are powerful.
Re: Qudits: The Real Future of Quantum Computing?
#9I know these guys understand quantum anything more than I ever will, but is having a larger radix specifically a practical way for increasing performance? If a binary device has a hardware fault, bits that are in error still have a 50% chance of being correct. It seems like a single component failing here would make this worse (five times worse if it wasn't quantum). I don't know whether or not the pros outweigh the…
A qudit expands the state space, sure. But now you need universal control over this system so I'm not sure you've bought anything. Also, a system will decay from energy level n to n-1 faster when n is larger... so like eh.
Re: Qudits: The Real Future of Quantum Computing?
#10You don't actually want qubits, you want an analog computer with differentiable signals. Most likely photonic. Qubits are a dead evolution branch. I've been recently exploring computational metamaterials for photonic computation. http://users.ece.utexas.edu/~aalu/research%20-%20page%203.ht... (there's quite a few papers on this but unfortunately they are all paywalled. Spoiler alert, they seem to be based entirely on…
This may be a rather naive question, as I've never seriously considered a modern take on analog computers with differentiable signals, but intuitively it's easy to see qualitative advantages in algorithms based on differential calculus, Fourier analysis, etc. What's less obvious to me is how such machines would do at fundamentally discrete tasks. For a trivially simple concrete example, consider the task of computing…
> What's less obvious to me is how such machines would do at fundamentally discrete tasks.
I might be talking out of my butt but fundamentally I think that eventually we'll realize that a lot of the currently supposedly solved problems might need to be rethought. I'm inclined to believe that hashing falls under these but don't quote me on that.
As for literature, I can't recommend anything in particular. But there's something on amazon, arxiv etc. I'm not sure what part of this are you most interested in, I guess what you are talking about could be complexity?