[darsnack]

You are right about qutrits. Essentially, anything that exhibits quantum behavior is in infinitely many levels at once. It is our ability to measure discrete levels that leads to qubits, qutrits, etc. Generally speaking, the more levels the more unreliable the measurement.

As for Monte Carlo with respect to ML, are you referring to the “random walk” aspect when you say “use randomness.” This refers to the fact that the method samples one possible sequence of events and uses that sequence to update its model. As opposed to dynamic programming methods where the value of all possible sequences is estimated and used to update the model. Not sure that the randomness from quantum is useful here. Where it could be useful is to encourage exploration of the state space. So, normally Monte Carlo methods have various tricks to make sure every sequence is sampled infinitely many times. These tricks could be implemented by the error in quantum computing, but I don’t know enough about the field to really be sure of that. And of course it would help in that you could compute the results of many sequences at once which is critical bottleneck in dynamic programming.

[sixdimensional]

You're thinking in the same direction I am.. but I am definitely too far out of my knowledge level to take it much further without sitting down and really working hard on the problem. I'm going off the deep end here into what is probably pop science or science fiction and might not make any sense at all. I wish I understood more about this in detail.

I just always observed that quantum computing was interesting because certain states seemed unpredictable, which is why scientists studying quantum computing struggle when they don't get the expected states. If that unpredictability is actually random due to the function of quantum states, then it would provide a unique tool since true randomness really is not easily found.

And the ability to potentially generate many random/expected states quickly - yes, I was thinking about being able to generate/explore the space faster. Or to use that in logical processing somehow as an advantage.

I have always been fascinated by a simple concept in logic - "maybe". I'm borrowing from "fuzzy logic" a bit here and twisting it the way I wish for my thought process. If one formalizes "yes","no" and "maybe" (i.e. 0, 1 and something else), can one then make complex logical statements involving probabilities, randomness, etc. by exploiting the "maybe" case.

For example, if I have one of those magic eight-balls, but it represents quantum states - if I can clearly detect 1 or 0 but sometimes I get a 0.5 or a 0.8 unpredictably, how could I use that to my advantage...

Another line of thinking for this is neural networks, another thing which I wonder could be accelerated by quantum computing. If I make the analogy that, as a human with a brain that is like a neural network, when I am faced with inputs that I have not processed before, my brain has to come up with a solution to process it from my various experiences and it might product an output that makes sense, or it might not.

Knowing my own fallibility here, if I know that sometimes I am faced with inputs that I don't know how to process or didn't expect, I can find a logic for how to handle those situations that might let me work around them.

[1] https://en.wikipedia.org/wiki/Fuzzy_logic

[virattara]

What I think, measurement of a qubit returns 0 or 1 always. The unpredictability is in the part where you were expecting 0 and got 1 because of noise I guess. But the measurement of a qubit in superposition does result in a purely random result.