the quantum complexity of time travel scott aaronson (mit)
TRANSCRIPT
Things we never see…
Warp drive Perpetuum mobile
GOLDBACH CONJECTURE: TRUE
NEXT QUESTION
Übercomputer
Does the absence of these devices tell us anything fundamental about physics?
In the first two cases, the answer is obvious
My view: It’s also obvious in the third case
My Research Interest:What We Can’t Do With Computers We Don’t Have
The Limits of Quantum Computers:
• Could quantum computers solve NP-complete problems in polynomial time?
• Could they break any cryptographic code (not just RSA)?
Evidence strongly suggests no
Most people don’t know this
What about analog computers, or quantum gravity computers, or…
This talk: Closed timelike curve
computers
Everyone’s first idea for a time travel computer: Do an arbitrarily long computation, then send the answer back in time to before you started
THIS DOES NOT WORK
Why not?
• Ignores the Grandfather Paradox
• Doesn’t take into account the computation you’ll have to do after getting the answer
Even in this bizarre setting, still need to quantify computational resources
David Deutsch’s Model
A closed timelike curve (CTC) is a computational resource that, given a function f, immediately finds a fixed point of f—that is, an x such that f(x)=x
Problem: Not every f has a fixed point!
But there’s always a distribution D such that f(D)=D
Probabilistic Resolution of the Grandfather Paradox- You’re born with ½ probability- If you’re born, you back and kill your grandfather- Hence you’re born with ½ probability
Question: What problems can be solved in this model?
Theorem: Exactly those problems solvable on a classical computer with a polynomial amount of memory, but possibly exponential time (the class PSPACE)
In other words, CTC’s make space and time equivalent as computational resources
You get to specify a polynomial time computation C, mapping n-bit strings to n-bit strings
Then Nature adversarially chooses a fixed point of the computation: a distribution D such C(D)=D
You get a sample from D
The computational model
The Nontrivial Question
What if we can perform a polynomial-time quantum computation inside the closed timelike curve?
Then certainly we can at least do PSPACE, since quantum computers can always simulate classical ones
But can we do more than PSPACE?
Three years ago I raised this as an open problem
Recently John Watrous and I managed to give a negative answer: if closed timelike curves exist, then quantum computers are no more powerful than classical ones
How did we show this?
Furthermore, we can compute P exactly in PSPACE, using Csanky’s NC2 algorithm for matrix inversion
111lim:
zMIzP
zSolution: Let
PMzzMIzMMPz
22
11lim
Then by Taylor expansion,
Hence P projects onto the fixed points of M
Let vec() be a “vectorization” of . We can reduce the problem to the following: given an (implicit) 22n22n matrix M, prepare a state in BQPSPACE such that
vecvec M
Conclusions
If closed timelike curves existed, then besides all the other strange implications, we could efficiently solve PSPACE-complete computational problems
For me, this is just additional evidence that closed timelike curves don’t exist
And yet, even in a world with closed timelike curves, we still wouldn’t have infinite computational power
Also, throwing quantum computing into the mix wouldn’t increase that power any further