0

I recently learned of a technique known as "block-encoding" which embeds any $M \times N$ matrix into a unitary matrix, given that the spectral norm is at most $1$. This type of result is pretty astounding--- but with this being able to work, I have had a thought.

Is there a method to embed a nonlinear operator into a quantum circuit? This does not necessarily have to be constrained by the math around quantum circuits, but it can be a physical process perhaps?

I don't see how this is non-reversible, as this type of nonlinear operator can be embedded into a reversible operator, similar to the block encoding technique.

Loic Stoic
  • 423
  • 1
  • 10
  • I think that you should explain a bit more what you want to do in order to get answers. What do you exactly mean by "embed a nonlinear operator into a quantum circuit". I believe it could mean many different things depending on the context. For your last paragraph: you can encode non-reversible operation in a reversible circuit. The trick is to ignore some outputs. Look at this for instance: https://quantumcomputing.stackexchange.com/questions/5829/implementing-classical-and-gate-and-classical-or-gate-with-a-quantum-circuit – Marco Fellous-Asiani Nov 03 '22 at 14:13
  • But you can't represent arbitrary non-linear maps by matrices. See this post for a discussion along these lines: https://math.stackexchange.com/questions/450/can-non-linear-transformations-be-represented-as-transformation-matrices – Ghost-of-PPPF Nov 03 '22 at 17:29

0 Answers0