1+1=3

Friday, December 21, 2007

I am visiting the Institute for Quantum Computing in Waterloo, Canada. It's such a great place to work, with both a guy with dreams and a guy with money to make them happen on the driving board. It's amazing what they've achieved, and what they want to in the future. It seems to me the best environment I have ever seen to work on quantum information. Only problem is...it's in Waterloo. I am thinking that maybe for a faculty position it makes sense to come here, or maybe for a postdoc without family. Otherwise, it would take a very sacrificed spouse (especially if he/she is not in science).

Saturday, December 15, 2007

A few weeks ago I was in Erice, Sicily, attending the Noise and Complexity at the Quantum Scale conference. The place was very, very nice, only a little cold to my taste.

This is a recollection of selected talks (the ones I took notes because I was interested).

H. Wiseman talked about quantum noise. His definition goes something like "I have a master equation, I add noise and obtain a sctochastic master equation". Each realization of noise is an unravelling of the master equation, and the noise comes mainly from measurement, it looks like it could be classical but no because of non-locality. This took him to discuss non-locality, and the three degrees of it they can identify (in order of strength): 3) Nonseparable, 2) steerable (can be used for EPR experiments), and 1) Bell non-local.

Peter Hanggi talked basically about their studies of Landau-Zaner effect with environment. They have a spin coupled to a bath of harmonic oscillators. For some reason, they think that when you have ZZ coupling to your environment you don't affect the Landau-Zener transition probability. But of course, in that model there is only dephasing so why it should be affected! I can prove it if the bath has no self-Hamiltonian.

G.M. D'Ariano gave a small talk that started good: what is the quantum equivalent of learning, such as in neural networks?

Franco Nori is too energetic. He flashed 289 slides in his 30 minutes, about I don't know how many experiments.

Barry Sanders gave a very interesting talk about Trotter-Suzuki decomposition for quantum simulations. They have been able to obtain the prefactor on the error of the TS approximation, which is also valid for my classical use of it. With this they obtained a precise bound on the number of gates you need for a simulation. There was also something that used only log-star steps but I can't remember what it was, I think it was the decomposition of the Hamiltonian without knowing it a priori.

H. Briegel started big talking about consciousness but then ended up with a toy model of a gas of classical particles with spin. The interesting part is that he had them in a box with two openings and was studying the entanglement in the steady (flow) state. There seemed to be a non-equilibrium transition in the entanglement as a function of the flow rate.

Zeilinger gave such a great talk, full of historical remarks, insight, and perspective.

E. Bagan (Barcelona) talked about quantum correlated measurements, the Chernoff bound and coin tossing for state discriminations. I liked this result: tr (B(A^t-B^t)>=0, for t>0 and A-B>0.

Casati gave a mixed talk about the Loschmidt echo, where I objected to his view that everything happens before the Ehrenfest time, and then he said that a quantum simulator of a chaotic system is not sensitive to perturbations (no Lyapunov regime). I have to check that because I felt there was a problem in his argument.

Plenio gave a nice talk about channel capacity near quantum criticality. We should invite him to the quantum lunch. The capacity by the way goes up to one at the critical point.

Finally, Seth Lloyd talked about quantum illumination. Still nobody laughs at his quantum illuminati joke. The scheme is to take an entangled pair of photons, send one to the detector and another one to a mirror, and then wait for a click on the entangled measurement. He sees an increase in efficiency, but I don't see at all why one of the photons goes up and the other down. From his calculations, it is equivalent to have both photons up. In fact, if he takes his object to be something different than a mirror, this would be much better since they could just gain an overall phase instead of the up-photon getting his phase changed and thus the whole measurement ruined. I have to talk to him about this.

This is a recollection of selected talks (the ones I took notes because I was interested).

H. Wiseman talked about quantum noise. His definition goes something like "I have a master equation, I add noise and obtain a sctochastic master equation". Each realization of noise is an unravelling of the master equation, and the noise comes mainly from measurement, it looks like it could be classical but no because of non-locality. This took him to discuss non-locality, and the three degrees of it they can identify (in order of strength): 3) Nonseparable, 2) steerable (can be used for EPR experiments), and 1) Bell non-local.

Peter Hanggi talked basically about their studies of Landau-Zaner effect with environment. They have a spin coupled to a bath of harmonic oscillators. For some reason, they think that when you have ZZ coupling to your environment you don't affect the Landau-Zener transition probability. But of course, in that model there is only dephasing so why it should be affected! I can prove it if the bath has no self-Hamiltonian.

G.M. D'Ariano gave a small talk that started good: what is the quantum equivalent of learning, such as in neural networks?

Franco Nori is too energetic. He flashed 289 slides in his 30 minutes, about I don't know how many experiments.

Barry Sanders gave a very interesting talk about Trotter-Suzuki decomposition for quantum simulations. They have been able to obtain the prefactor on the error of the TS approximation, which is also valid for my classical use of it. With this they obtained a precise bound on the number of gates you need for a simulation. There was also something that used only log-star steps but I can't remember what it was, I think it was the decomposition of the Hamiltonian without knowing it a priori.

H. Briegel started big talking about consciousness but then ended up with a toy model of a gas of classical particles with spin. The interesting part is that he had them in a box with two openings and was studying the entanglement in the steady (flow) state. There seemed to be a non-equilibrium transition in the entanglement as a function of the flow rate.

Zeilinger gave such a great talk, full of historical remarks, insight, and perspective.

E. Bagan (Barcelona) talked about quantum correlated measurements, the Chernoff bound and coin tossing for state discriminations. I liked this result: tr (B(A^t-B^t)>=0, for t>0 and A-B>0.

Casati gave a mixed talk about the Loschmidt echo, where I objected to his view that everything happens before the Ehrenfest time, and then he said that a quantum simulator of a chaotic system is not sensitive to perturbations (no Lyapunov regime). I have to check that because I felt there was a problem in his argument.

Plenio gave a nice talk about channel capacity near quantum criticality. We should invite him to the quantum lunch. The capacity by the way goes up to one at the critical point.

Finally, Seth Lloyd talked about quantum illumination. Still nobody laughs at his quantum illuminati joke. The scheme is to take an entangled pair of photons, send one to the detector and another one to a mirror, and then wait for a click on the entangled measurement. He sees an increase in efficiency, but I don't see at all why one of the photons goes up and the other down. From his calculations, it is equivalent to have both photons up. In fact, if he takes his object to be something different than a mirror, this would be much better since they could just gain an overall phase instead of the up-photon getting his phase changed and thus the whole measurement ruined. I have to talk to him about this.

Tuesday, December 11, 2007

A lot of work near the end of the year (always feeling it's some a kind of landmark). Two or three projects that I wanted to have finished are...well, close enough. One of them might receive all the attention. I'll be presenting a poster on wednesday 11th at LANL -- for different reasons, that's three posters this year after my promise not to make any more posters. --, and next week I'll be giving a seminar at IQC in Waterloo. I hope the weather is at least good enough to let me fly without delays. I almost finished applying to jobs, which took most of my last two months. For next year, I got a strange start with my parents visiting for all of january. I would like to finish my MPI implementation of Trotter-Suzuki for BECs. That would let me do really cool and big simulations. I am also daydreaming of threading the application to use the eight cores of my Mac Pro...

For large values of 1.

05/01/2006 - 06/01/2006

06/01/2006 - 07/01/2006

07/01/2006 - 08/01/2006

08/01/2006 - 09/01/2006

09/01/2006 - 10/01/2006

10/01/2006 - 11/01/2006

11/01/2006 - 12/01/2006

12/01/2006 - 01/01/2007

02/01/2007 - 03/01/2007

03/01/2007 - 04/01/2007

05/01/2007 - 06/01/2007

08/01/2007 - 09/01/2007

11/01/2007 - 12/01/2007

12/01/2007 - 01/01/2008

01/01/2008 - 02/01/2008

02/01/2008 - 03/01/2008

04/01/2008 - 05/01/2008

05/01/2008 - 06/01/2008

06/01/2008 - 07/01/2008

07/01/2008 - 08/01/2008

08/01/2008 - 09/01/2008

04/01/2009 - 05/01/2009

12/01/2010 - 01/01/2011

01/01/2011 - 02/01/2011