Quantum Bayesian Networks

January 16, 2012

Bayesian Networks + Information Theory = Papa-the-Camel’s Book

Filed under: Uncategorized — rrtucci @ 5:59 pm

My religious epiphany for the week:
For those of you who, like me, are fans of both

  • Bayesian Networks (a divine theory, also known by many other names, just like God/Yahweh/Allah etc. is) and

  • Shannon Information Theory (SIT) (another divinely inspired theory),

here is an excellent book that merges these two fascinating topics. A preliminary draft of the book can be downloaded from ArXiv.

Lecture Notes on Network Information Theory
Abbas El Gamal, Young-Han Kim

(I am told by my impeccable internet sources that “El Gamal” means “The Camel” in Arabic. “Abbas” means “papa”. It does remind me of capacity. Information Capacity Regions is a fascinating subject which the El Gamal/Kim book often discusses.)

The arxiv version of the book is quite adequate for many purposes, but a more polished version has recently been published (in fact, published just today! No stale news in this blog!) by Cambridge University Press. 714 pages of goodness. Amazon link to book

One of my lifelong goals is to build a quantum version of B.Nets and SIT, and to apply it to physics. I’m of course not the only one, or even the first one, to arrive at this idea. I see that Mark Wilde, a rising luminary of quantum information theory, already cites the El Gamal/Kim book in several of his arxiv papers. Good find Mark!

Abbas El Gamal is a professor at Stanford University. Stanford is one of the CHIMPS of quantum computing, meaning they haven’t done much in that area. On the other hand, Stanford is the home of Cover and Thomas, the authors of one of the canonical books on SIT, and also the home of Daphne Koller, coauthor of one of the canonical books about B nets. I think eventually Stanford will learn how to add simple things like
quantum + B.Nets + SIT + physics.

By the way, Stanford is also home of the indescribable Lenny Susskind, String Theorist extraordinaire, who, if my prediction for 2012 is right, will soon be singing Q-Comp, Q-Comp!. Once already, a few years ago, Lenny invited Scott Aaronson to speak at one of his String Theory shindigs, so Lenny has already shown some potential for conversion to our QC faith.

January 10, 2012

Unethical or Really Dumb (or both) Scientists from University of Adelaide “Rediscover” My Version of Grover’s Algorithm

Filed under: Uncategorized — rrtucci @ 2:56 am

This new paper

An improved formalism for the Grover search algorithm,
by James M. Chappell (1,2) , M. A. Lohe (2), Lorenz von Smekal (3), Azhar Iqbal (1) and Derek Abbott (1)

  1. School of Electrical and Electronic Engineering, University of Adelaide 5005, Australia
  2. School of Chemistry and Physics, University of Adelaide, South Australia 5005, Australia
  3. Institut f¨ur Kernphysik, Technische Universit¨at Darmstadt, Schlossgartenstraße 9, 64289 Darmstadt, Germany

“rediscovers” the fixed-point version of Grover’s algorithm that I invented and presented in the following paper:

An Adaptive, Fixed-Point Version of Grover’s Algorithm,
by Robert R. Tucci

The Chappell et al. paper has 24 references but does not refer to my paper, even though their paper and mine are eerily similar. Compare them yourself. With the excellent Google and ArXiv search engines, I would say there is zero probability that none of its five authors knew about my paper before they wrote theirs. (I have also written multiple posts in this blog describing my algorithm and linking to my arxiv paper about it, so nobody can claim that it was kept well hidden)

Addendum Jan 16, 2012:
Now that my emotions have calmed down, I was able to look at the Chappell et al paper more carefully. Our algorithms are more different than I initially thought. I’m not sure yet if one algorithm is faster or more accurate than the other. They may be about the same, or they may not be. More analysis will be required to answer this. I still think Chappell et al should have mentioned my paper. I would have done so if I had been in their shoes. After all, my paper was published 2 years earlier than theirs, and it solves the same problem, albeit using a different strategy.

January 6, 2012

Of Bananas and String Theorists

Filed under: Uncategorized — rrtucci @ 5:57 pm

I figured out how to distill the message of the previous post into an ape picture. Ape pictures make difficult concepts more intuitively understandable. They provide much guidance for physicists.

January 1, 2012

My Prediction for 2012: Quantum Computerists Will Occupy String Theory Bastions Like Princeton’s IAS

Filed under: Uncategorized — rrtucci @ 4:13 am

Quantum Computer Programmers Too

“String theorists don’t make predictions, they make excuses” -Feynman (1918-1988)

In America during the 1960’s, the hippies and draft dodgers rose up against the powers that be. At present, a similar phenomenon is sweeping through many Islamic countries (the so-called Arab Spring), and America (the Occupy Wall Street protest movement). It seems that every 50 years or so, the young rise up against the ruling class elites and abusers of power.

I predict that in 2012, a new protest movement will bubble up from that boiling cauldron of unrest that is Academia, and start a new String Theory revolution.

I, Nostradamucci, see the future clearly. The end times. A world in turmoil. Four horsemen. QC barbarians at the String Theory ramparts. En masse occupation of IAS by quantum computerists. Mass defections from the String Theory ranks to the QC ranks. A Great Cheapening of the once venerable String Theory brand. The QC Winter. The end of innocence.

There have been some recent, very foreboding developments at the crossroads between high energy physics and quantum computing, that give my prediction an air of inevitability. Here are some of those ominous signs:

  • In a previous blog post of mine,

    Set a Thief to Catch a Thief

    I gave a brief historical review of the idea of simulating quantum mechanical systems using a QC.

    There is a steadily increasing logjam of quantum mechanical theories (examples can be found in the fields of quantum gravity in general, String Theory in particular, relativistic quantum field theory, condensed matter physics, etc. ) that resist analysis in their strong and intermediate coupling regimes, regimes where an expansion in powers of the coupling constant fails to converge. So far nobody knows how to milk such theories very well.

    Perhaps if we could simulate such theories on some kind of computer, we could compare their predictions with laboratory measurements. We might then be able to discard or modify or refine those theories. We might also be able to simplify such theories to their sweet spot, the spot where they are simple enough that we can figure out how to solve them analytically, yet complicated enough that they still make useful predictions which agree reasonably well with the laboratory data.

    A QC would be the ideal device for simulating such currently intractable quantum theories. Feynman himself was the first to point out that using a QC instead of a classical computer to simulate a quantum system allows one to perform such simulations faster by a factor exp(n), where n is the number of bits in the input data. Also, no need for iffy Wick-y rotations. And QCs can simulate both the strong and weak coupling regimes of a theory with the same software code and with the greatest of ease.

  • Recently, some researchers have proposed in the following paper, a method similar to the methods discussed in the above Thieves post, but this time for using a QC to simulate the phi^4 relativistic quantum field theory in 3+1 dimensions. They have worked out some of the nitty-gritty details of Feynman’s original dream:

    Quantum Computation of Scattering in Scalar Quantum Field Theories
    Stephen P. Jordan, Keith S. M. Lee, John Preskill

    Papers that give methods for doing QC simulations of quantum field theories that also include fermions and gauge fields, are no doubt being written as we speak. Much work remains to be done in this area. All the details of how one would simulate a String Theory on a QC remain undiscovered yet. Maybe you will be the one who finds them.

  • Lattice Gauge Theories were invented circa 1974 by Ken Wilson and others. It has taken scientists many decades of grueling work, using Monte Carlo techniques on classical supercomputers, to simulate lattice QCD (Quantum Chromodynamics), and to derive the mass spectra predicted by it. Imagine if QCs could do all that work in seconds.

    On classical computers, to reduce the computational-resource requirements to a non-prohibitive size, one is often forced to butcher the fermionic degrees of freedom of the lattice gauge theory by using ugly, poorly understood approximations. This would not be necessary if the simulations were done on a QC.

    QCs can do Monte Carlo much faster than classical computers, a fact which I’ve explained in many previous posts in this blog, and which I exploit in my own QC Gibbs sampling software called Quibbs.

  • Recently there have been some paltry attempts to use disgustingly classical supercomputers to try to squeeze out some straight answers from String Theory. Quantum computers could do such analyses much more quickly, accurately and painlessly.

  • Machiavelli speaks: Ascolta, mio principe (Listen, my Prince): String theory is a hard subject to sell to the general public. No verified quantitative predictions so far. No practical applications. But if string theorists start selling themselves by saying that we are helping to develop quantum computers at the same time that we explore our theory, they might get some new buyers. Even the military and DARPA, which don’t normally fund string theorists, might bite.

  • An anonymous quantum computerists, probably an ex-physicist working in Wall Street, recently sent me the following wordy, demented manifesto for a protest movement that he is trying to start:

    We are the 99% who are not String Theorists. Let’s now define the set of people who are not String Theorists as being Quantum Computerists, just like Lilliput had its Big Enders and Small Enders. We Quantum Computerists believe that String Theorists who work at primo places like Princeton’s IAS (the Institute for Advanced Study, where Ed Witten works, where Einstein and Godel once worked, a place rated by Sidney Coleman as the most boring town he has ever lived in) control too much of the wealth (in funding booty, post doc slaves and public imagination) of physics. And so far String Theorists have squandered that wealth and patrimony.

    Let’s face it, investments in String Theory were a huge gamble from the very beginning—I mean, even at its christening, Feynman didn’t give his blessing to John Schwarz’s baby, but Feynman certainly did give his blessing to his own baby, quantum computing.

    String Theorists have failed miserably to balance their books (no good quantitative agreement between the two sides of the ledger, theory and experiment). But they are not being held accountable for their poor book-keeping practices. They continue to be bailed out by the government because their business is considered too big to fail, according to Peter Woit.

    Even after the Great Recession caused by multiverse ideas, most String Theorists still deny that they ever did anything wrong. They continue to do business as usual, steadfast in their Enron-esque opinion that they are the smartest people in the room, and, therefore, they can do no wrong. Well, I’ve got news for them, some quantum computerists (most notably the complexity theorists) think they are even smarter. (note from rrtucci: Leave me out of it. I belong to the 99% of the 99% who are as blissfully dumb as Ed Witten’s hamster.)

    We quantum computerists are not against laissez faire capitalism. That would be like being against truth, justice and the American way. So we are not against String Theory per se either, or against any other lesser theories of quantum gravity. What we deplore is that the gains made by String Theorists in the past few decades are not trickling down to us.

    Let’s face it. String Theory is as transparent as a Credit Default Swap. Perhaps quantum computerists can make String Theory more transparent by simulating various facets of it on a quantum computer?

Create a free website or blog at WordPress.com.

%d bloggers like this: