QIP '10 Liveblogging - Day 0.5
Loïck
on Monday, January 18 2010, 14:45
Live-blogging
QIP = PSPACE
Rahul Jain, Zhengfeng Jin Sarvagya Upadhyay, John Watrous — arXiv:0907.4737
Known facts:
- IP = PSPACE
- QIP = QIP(3) [Watrous99, KitaevWatrous00]
- QMA = QIP(1) ⊆ PSPACE
- QIP(2) ⊆ PSPACE [JainUpadhyayWatrous09] (uses SDP)
- QIP = QMAM [MarriotWatrous05]
Need to show that QMAM ⊆ PSPACE. The key is to write the verifier’s Maximum Acceptance Probability (MAP) as a SDP and find feasible points. Pick some point of the primal. If they are not, they are (up to a projection) dual feasible points. Then use the MWU in order to improve the solution of the dual. Try log(size of the problem) times.
(and I did not get how to have a good solution of the primal from them) (strong duality?)
QIP = IP : Quantum Interactive Proofs are not more powerful than classical interactive proofs.
Random numbers certified by Bell’s theorem
Antonio Acín, Antoine Boyer and Stefano Pironio — arXiv:0911.3427
How can you trust your random number generators? Seriously: you have to be paranoid about your RNG. Do it yourself! You just need a device that violates Bell inequalities and extract the quantum randomness in it.
Protocol: you have two devices: you feed them with N join random bits and you compute the estimator of the violation of Bell inequalities with their outputs.
Result: A bound on the output entropy as a function of the violation of Bell inequalities: H∞ ≥ N x f(I-ε) with probability 1-δ. where I is the estimator of the violation, ε and δ are small compared to N and f is a function.
Assumptions
- your devices take bits on input and outputs bits (can have memory)
- your two devices cannot communicate (but they can share entanglement)
- the adversary has bounded quantum memory (it is not composable (yet!))
The output string is not necessary uniform. You need to do randomness extraction to actually have useful randomness.
Adiabatic gate teleportation
Dave Bacon and Steve Flammia — arXiv:0905.0901 and arXiv:0912.2098
Steve is laughing at computer scientists that don’t like integrals.
Adiabatic teleportation: you make teleportion in an adiabatic way…
Adiabatic gate teleportation: you start with a qubit |ψ > and at the end of the teleportation you get U|ψ>
For a 1 qubit gate, you just need to perform a rotation on one qubit of your EPR pair before the usual teleportation. For 2 qubit gates, it is more complicated, you need to use a gadget and make 2 teleportations.
Comments
Monday, January 18 2010, 15:12
If you want the subset to appear properly, using the proper unicode characters, you can copy'n'paste the following:
QIP = QIP(3)%%%
QIP ⊆ EXP
QIP(1) aka QMA ⊆ PP ⊆ PSPACE
QIP(2) ⊆ PSPACE
Monday, January 18 2010, 15:18
Danke !
Monday, January 18 2010, 15:22
By the way, http://tlt.its.psu.edu/suggestions/... says that ther is a standard html nabe for the symbol : ⊆ . It might be easier then cut'n paste.
Monday, January 18 2010, 15:24
I don't give a damn to HTML entities: UTF8 is my friend!
Monday, January 18 2010, 22:53
Wow... what a COOL font your site uses...
I wonder what it will look like under linux or windows, but on Mac it is BEAUTIFUL.
Suggestion fro QIP live blogging....
Perhaps you guys can use a shared notebook to collectively take notes and shoot info at onlookers?
I created one for that purpose on etherpad :
http://etherpad.com/YiYvglzGAS
A+
Tuesday, January 19 2010, 09:29
I use the font Diavlo. The rendering does not depend of your OS but on the ability to your browser to use downloadable fonts. I won't use etherpad, but you're free to copy/paste by blog posts…