Wednesday, March 09, 2005

WMAP and cosmic strings

This is just a small note about the cosmic strings; the previous article about this topic described Mark Jackson's seminar.

Quantoken asked what's new with the observed cosmic strings - his or her personal theory is that CSL-1 "was really just bird shit causing the original telescope to record a slightly distorted image".

The next debater immediately pointed out a new article astro-ph/0503120:

The article first of all confirms that there are no cosmic strings seen in the microwave background whose tension would exceed a certain bound and whose velocity is smaller than 0.7c or so. Then they focus on the neighborhood of CSL-1, the double galaxy conjectured to be a result of gravitational lensing by a cosmic string. Yes, they use the same coordinates from Mark Jackson that I published in the previous blog article.

The result?
  • They observe the discontinuity of the microwave background temperature near the cosmic string with *2 sigma* significance which may or may not be a sign of the cosmic string. The whole CSL-1 double galaxy fits into one pixel of or so, so they have to use 4 pixels around, and the precision is not great. PLANCK is expected to be twice as sensitive. The cosmic string would have to move by velocity higher than 0.94c to fit the data.
Henry Tye at Harvard

On Thursday, Henry Tye was giving a nice talk at the Duality Seminar about the cosmic strings. It included a lot of new interesting information about the simulation of cosmic string networks, about the WMAP - cosmic string relation described above, and so forth. He also said that the number of cosmic strings from an Abelian i.e. U(1) Higgs model in the visible Universe should be 40 plus minus a few. If it's not fourty, the strings must be strings from string theory, some people say! ;-)

Many well-known physicists - like Cumrun Vafa, Hiroshi Ooguri, Juan Maldacena (who was speaking at MIT which I unfortunately missed), Henry Tye, Jacques Distler (interesting visitors, aren't they?!) - discussed the nature of the Hartle-Hawking state. Incidentally, one of the results is that Jacques' viewpoint that the conformal factor makes Euclidean quantum gravity inconsistent was largerly debunked, if Jacques allows me to say. We seemed to agree that the problems can't survive in anti de Sitter space where the things seem to work, and it has also been explained that the reason why it works is not because of the negative cosmological constant term (one would in fact need a positive cosmological constant if this term should be able to cure the unboundedness of the action from below).

Now we returned from the dinner with the speaker at Sushi Cafe (Hirosi, Henry, Jacques were the more senior people). The debates were very interesting - cosmology, string theory, outsourcing of string theorists to India, increasing capacity of memory chips, microprocessors, hard disks, and so forth.

Larry Summers and the physicists

Also, before the dinner, between 6:15 pm and 7:00 pm, Larry Summers debated roughly 100+ physicists in the physics library. Lisa Randall introduced the president - and I think that she was pretty balanced. Everyone I talked to has been satisfied with Summers' answers, and everyone said he was nice and so on. No doubt, a hypothetical dissatisfaction with the president is partially another example of the science wars - the disagreement between hard sciences on one side and soft sciences on the other side; of course, economics is counted as a hard science, especially in Larry's edition.

Let me mention several highlights. Someone was asking whom should a student or scholar try to contact if she or he faces some grievances or discrimination of any kind - someone outside the "political hierarchy" of the school. Summers explained that there exists the ombudsperson at Harvard. An instant vote proposed by Summers immediately revealed that 2/3 of the people did not know that the ombudsperson existed (I was among the ignorant two thirds). Another girl who spoke afterwards showed that even the important people at Harvard did not know that the ombudsperson existed.

Although Summers has said many sentences that were more politically correct and polite than informative ;-), he also stated many important things that were not affected by PC. He defined his vision of Harvard and sciences at Harvard, and he insisted that the policies must be effective, Harvard must be the leader, and other schools will undoubtedly follow this instutution every time Harvard invents something that works (Yale's support of the students from poor families is a recent example). Other questions led Summers to explain how the hiring decisions are made; how the support of the poorer scholars and those with children is working now and how it will work in the future.

At the end, Summers has pointed out that the men (and women) who attended the meeting are likely to be a part of the solution. In other words, the likely parts of the problem are those who did not attend. The last thing I have to mention - and the relation of this sentence with the previous sentences is purely coincidental if any - is that Dan Fisher did not attend. ;-)


  1. I still can't get the picture of a moose running around with feyman's toy models for antlers:)I was going to try and make a picture of a moose doing such.

    What I wanted to say asa suggestion is that you should keep seperate the issue of cosmic strings and Larry Summers, as I understand the views of some that would comment there and some who would comment on cosmic strings?

    No takers?

    I wanted to respond to cosmic strings. We had to to see how these strings existed in context of supersymmetry?

  2. Lubos said: "They observe the discontinuity of the microwave background temperature near the cosmic string with *2 sigma* significance which may or may not be a sign of the cosmic string. The whole CSL-1 double galaxy fits into one pixel of WMAP or so, so they have to use 4 pixels around, and the precision is not great. PLANCK is expected to be twice as sensitive. The cosmic string would have to move by velocity higher than 0.94c to fit the data."

    That's another ghost chasing story! The so called CSL-1 feature, being 10 billion light years away, is only 2 arcseconds. The WMAP resolution is not very high at all. The paper listed 13 arc minutes as the pixel resolution. See the paper:

    I know this guy Edward Wright and had read his web site. He is a complete idiot unable to even count numbers using fingers!

    Top of page 6, he cited: "At 13 arc-minute resolution, the sky is divided "12x4^9 pixels". That looks very odd to me. Calculating from a square of 13 arc-minutes per pixel, you would thought for a whole sky of 4*PI solid angle, there's about 9x10^5 pixels, not "12x4^9". It is also an odd way, none-scientific way of expressing a quantity like "12x4^9". The scientific way would have been X * 10^Y.

    There are 10^11 to 10^12 galaxies in the whole universe. Divide it into 9x10^5 pixels for the whole sky, each pixel will contain 10^5 to 10^6 galaxies, each as big as CSL-1.

    So CSL-1 is just one out of 10^5 to 10^6 galaxies whose photons fall onto the same pixel in the WMAP dataset. Even if there is the so called cosmic string on CSL-1, it probably would't result in any recognizable feature after all photons from all 10^5 to 10^6 galaxies mix into data of just one pixel.

    Plus how significant is a 2 sigma random fluctuation? The first letter of Lubos's last name happen to be the same as "M-theory". That's probably a 3.5 sigma significance:-)

    Why I think this guy is stupid? On his web page where he argue against star light as the cause of CMB, he said Eddington's argument is only good within a galaxy. Well, not only he is wrong on this assertion. But it is also stupid for some one to write those words in a little office, and totally forget he sits in an office on the earth, which IS completely within a giant galaxy called milky way!!!


  3. "He also said that the number of cosmic strings from an Abelian i.e. U(1) Higgs model in the visible Universe should be 40 plus minus a few."

    Is this a typo or a joke? A string can couple magnetically to a scalar in 4d. There are 70 (seventy) compact scalars in 4 dimensions, and the same number of different cosmic strings.

    I did attend Maldacena's talk at MIT. As always his work is kind of cool. I don't have much time right now, but here is the punch line of his talk.

    Starting from N=4 SYM with global U(1) \times U(1) (non-R-) symmetry and its gravity dual, he constructs a class of exactly marginal deformations that break to N=4 to N=1 SUSY. These deformations can be carried out explicitly on *both* sides, gauge and gravity.

    The N=4 SYM has, in N=1 language, 3 chiral multiplets with Lagrangian

    L = Tr [ \partial \phi_i \partial \phi_i + \phi_1 \phi_2 \phi_3 - \phi_1 \phi_3 \phi_2 ]

    Now introduce a star product

    \phi_i * \phi_j = exp ( i \pi [\theta_i^1 \theta_j^2 - \theta_i^2 \theta_j^1] ) \phi_i * \phi_j

    The \theta^a (a=1,2) are the charges of the global U(1) \times U(1). This resembles the star-product of Seiberg-Witten non-commutative geometry, but it isn't quite the same. The difference is that here the phases depend on "phase space" indices rather than world volume coo's. At any rate, it does amount to putting \phi_i, \phi_j on a non-commutative torus. Use this * product, instead of the ordinary one, to define the Lagrangian. Choose phases \theta_i^a such that the kinetic term remains invariant and the two terms in the potential pick up a relative phase. This is the exactly marginal deformation that breaks SUSY to N=1. Suppose the relative phase is \gamma. Introduce a complex parameter

    \tau = V + i B,

    where V is the volume of the torus and B is the B-field. After the deformation you find

    \tau \to \tau' = (1+\gamma\tau)^{-1} \tau

    This is pretty a pretty darn simple SL(2,R) tranformation, and that's what's so cool about it.

    On the gravity side, the U(1) \times U(1) corresponds to isometries of the metric. Thus there are no surprises and you can already guess how the metric changes.

  4. Do you mind sharing with us how you debunked Distler?

  5. Hi, sure. I think that I've already explained it mostly in the article.

    Jacques' point was that there is some infrared problem associated with the scale factor that makes the whole quantum gravity in Euclidean spacetime inconsistent.

    It's known that the scale factor has locally a negative kinetic term but Jacques' viewpoint was that this was an actual inconsistency.

    This can be seen not to be the case in Euclidean anti de Sitter space where the Euclidean AdS path integral is well-defined in terms of the correlators on the boundary of EAdS. In fact, it is the Euclidean version of the theory that we normally calculate with.

    So Jacques then argued that his problem is cured in Anti de Sitter space because the cosmological constant term (Lambda) beats the Einstein-Hilbert term (R) and makes the action bounded from below after all.

    The c.c. term goes like the volume while the EH term goes like volume/radius squared, and therefore it loses.

    However, if you look at the signs, you easily find out that the sign of these two terms is always opposite assuming that you choose the appropriate (for the sign of the cosmological constant) maximal space of a variable radius. Why? It's simply because there must be a stationary point - the sphere or the EAdS - where the sign flips.

    The overall sign is of either term is of course opposite in EAdS and sphere. If you look at any potential unboundedness of the EH action from below in the Euclidean space like sphere, it is clear that there will be an even bigger unboundedness from below in the cosmological term in the EAdS space.

    So if the problem of Jacques were cured somewhere, it would likely be in de Sitter space, not anti de Sitter space as he argues. But even anti de Sitter space works OK.

    The conclusion is that Jacques' problem with Euclidean gravity only exists if someone does things incorrectly.

  6. So why is the Euclidean path integral well defined even though the action is unbounded from below? Are there any mathemetical "tricks" involved?

  7. Why would anyone write a paper about a 2 sigma effect? Pointless.

    Even the celebrated 115GeV Higgs non-signal was always rather bigger than 2 sigma.

    I was at a conference recently where there was a nice talk on statistical evidence and inference which explained roughly why 95% of all 2 sigma 'discoveries' are false positives. Paul Shellard (who should know something about cosmic strings!) also complained at the waste of resources entailed by people writing and having to referee such papers.