Blogger claims BICEP2 team acknowledging possible error in discovery of evidence of gravitational waves

May 14, 2014 by Bob Yirka weblog

(Phys.org) —Adam Falkowski, a physicist working at CERN, on his Particle Physics Blog, is claiming that researchers on the BICEP2 team that uploaded a paper (First direct evidence of cosmic inflation) to the arXiv preprint server this past March have acknowledged to some in the science community that there may be a problem with their methodology. Members of the BICEP2 research team are denying Falkowski's claim, but the assertion has led to rumors on the Internet that the team may not have found evidence of cosmic inflation after all.

The Big Bang theory suggests the universe as we know it came to exist as the result of a giant explosion, approximately 14 billion years ago, followed by a rapid thrusting of material from the point of the explosion out into what we now observe as the universe around us. That thrusting has been dubbed the theory of —it describes the rapid expansion of the universe. Adherents contend that during the initial phase of cosmic inflation, would have been generated, and should be still visible today. Last March, the BICEP2 team claimed in their paper that they had found evidence of such gravitational waves, giving cosmic inflation theory a huge boost.

But now, some have suggested that there might have been a problem with the way the observations were made—a map the team used which was created by the ESA's Planck team might have been interpreted incorrectly. To spot gravitational waves, the team had to rule out other signals it received based on data from a variety of sources. One of those sources, the map created by the Planck team included several possible sources of light, but the researchers at BICEP2 thought it only charted dust or ashes from exploding stars. This bit of news leads to a little bit of doubt about the results the team found, and now places them at the mercy of an update of the map by the Planck team.

In the meantime, members of the BICEP2 team have been responding to the rumors and claim they are still confident in their results and that any gossip suggesting they have begun to doubt their work is wrong. Whether they believe their work is in jeopardy or not, is, of course, not really what's important—finding out if what they reported is correct is what matters, and that is going to take some time.

Explore further: Cosmologists cast doubt on inflation evidence

add to favorites email to friend print save as pdf

Related Stories

Cosmologists cast doubt on inflation evidence

Mar 25, 2014

It was just a week ago that the news blew through the scientific world like a storm: researchers from the BICEP2 project at the South Pole Telescope had detected unambiguous evidence of primordial gravitational waves in the cosmic microwave background, t ...

First direct evidence of cosmic inflation (Update)

Mar 17, 2014

(Phys.org) —Almost 14 billion years ago, the universe we inhabit burst into existence in an extraordinary event that initiated the Big Bang. In the first fleeting fraction of a second, the universe expanded ...

Have cosmologists lost their minds in the multiverse?

May 13, 2014

The recent BICEP2 observations – of swirls in the polarisation of the cosmic microwave background – have been proclaimed as many things, from evidence of the Big Bang and gravitational waves to something ...

Setting a trap for gravity waves

Mar 19, 2014

In 1996 Uros Seljak was a postdoc at Harvard, contemplating ways to extract information from the cosmic microwave background (CMB). The distribution of anisotropies, slight temperature differences, in the ...

Recommended for you

Supernova seen in two lights

19 minutes ago

(Phys.org) —The destructive results of a mighty supernova explosion reveal themselves in a delicate blend of infrared and X-ray light, as seen in this image from NASA's Spitzer Space Telescope and Chandra ...

Toothpaste fluorine formed in stars

22 hours ago

The fluorine that is found in products such as toothpaste was likely formed billions of years ago in now dead stars of the same type as our sun. This has been shown by astronomers at Lund University in Sweden, ...

Swirling electrons in the whirlpool galaxy

Aug 20, 2014

The whirlpool galaxy Messier 51 (M51) is seen from a distance of approximately 30 million light years. This galaxy appears almost face-on and displays a beautiful system of spiral arms.

A spectacular landscape of star formation

Aug 20, 2014

This image, captured by the Wide Field Imager at ESO's La Silla Observatory in Chile, shows two dramatic star formation regions in the Milky Way. The first, on the left, is dominated by the star cluster NGC ...

User comments : 14

Adjust slider to filter visible comments by rank

Display comments: newest first

indio007
1.5 / 5 (8) May 14, 2014
We really need to get these guys doing MRI and any other EM sensing.

They are miracle workers that can detect polarization out of 1/1000 = signal/noise on the low end.
IMP-9
5 / 5 (3) May 14, 2014
1/1000 = signal/noise

That's not true.
shavera
4.5 / 5 (8) May 14, 2014
actually a lot of physicists or their work do/does end up going into medical imaging and the like. Whether your specific comment was facetious or not, the work done by physicists does help modern technology, and it trains people who don't stay in physics to help build new modern technology as well.
antialias_physorg
5 / 5 (3) May 14, 2014
They are miracle workers that can detect polarization out of 1/1000 = signal/noise on the low end.


The SNR for the BICEP study was 1 in 10 million.



1/1000 = signal/noise


That's not true.

Only a very small percentage of protons actually contribute to a signal in an MRI. If I remember the lectures at uni correctly then it's less than 1 proton in 6000 that gets 'extra' aligned with the field (i.e. contributes to the signal). The rest is noise.
IMP-9
3.7 / 5 (3) May 14, 2014
Only a very small percentage of protons actually contribute to a signal in an MRI. If I remember the lectures at uni correctly then it's less than 1 proton in 6000 that gets 'extra' aligned with the field (i.e. contributes to the signal). The rest is noise.


No, the rest is background. Signal to noise ratios do not include background. It doesn't matter how much background there is, all that matters is the fluctuation. That is not SNR. If you have an SNR ratio of 1 in 10 million you haven't detected anything.
Tipped
1 / 5 (2) May 14, 2014
Only a very small percentage of protons actually contribute to a signal in an MRI. If I remember the lectures at uni correctly then it's less than 1 proton in 6000 that gets 'extra' aligned with the field (i.e. contributes to the signal). The rest is noise.


No, the rest is background. Signal to noise ratios do not include background. It doesn't matter how much background there is, all that matters is the fluctuation. That is not SNR. If you have an SNR ratio of 1 in 10 million you haven't detected anything.

Oh really? Background is nothing but noise as far as signal processors are concerned. In cases where we have a good idea of what the noise looks like or pieces of the noise, we can filter that out and significantly reduce the noise in the output, which is what I think you're talking about.
nevermark
3 / 5 (2) May 14, 2014
IMP-9, actually any signal-to-noise ratio can result in a detectable signal. As long as the detector is sensitive enough to pick up the signals contribution, then higher noise levels simply mean more data must be available to detect the signal.
Torbjorn_Larsson_OM
5 / 5 (3) May 14, 2014
It's an unsubstantiated rumor, and even if it would be true it doesn't add anything to the BICEP2 paper. (And that is what BICEP2 scientists note.) The latter paper relied on a lack of crosscorrelation, not on the subtraction of dust. The attempted subtraction decrease significance, but with the data at hand not sufficiently. The data disfavor r = 0 (no tensor modes from PGW) without dust, and with available dust data such as the current dust data from Planck.

I'm dissatisfied with Jester (AF's nym) for the first time.

@indio007: The PTE (here Probability To Exceed is the same as Signal-to-Noise Ratio) was ~1 to 10 million indeed, but that corresponds to 5.3 sigma. (Page 8 of the paper.)
IMP-9
5 / 5 (2) May 14, 2014
Oh really? Background is nothing but noise as far as signal processors are concerned.

Noise and background are two completely different things. It's very important you know the difference. You can happily have a signal that is arbitrary small against the background but if these size of the signal is smaller than the size of the fluctuations in output (what noise actually is) then you can't claim to have detected anything. You need to do things to increase the SNR.

actually any signal-to-noise ratio can result in a detectable signal. As long as the detector is sensitive enough to pick up the signals contribution, then higher noise levels simply mean more data must be available to detect the signal.

No. Integrating is one way of increasing SNR. You aren't picking out a signal at small SNR you averaging out the noise to improve the SNR. Noise is characterised by the standard deviation of the fluctuations, you wouldn't claim you could detect a signal at 0.0001 sigma.
Rustybolts
2 / 5 (4) May 15, 2014
Bad theory to be working on the first place.
betterexists
1 / 5 (3) May 15, 2014
ERROR Not only in this. In Others Too!
/other/2014/4120fusion_crushed_usadothtml on larouchepubdotcom says:
"Nuclear fusion has been at Mankind's Fingertips for Decades; Could Already be providing the World with Virtually Unlimited Energy; it has been deliberately suppressed under a top-down Imperial policy"
betterexists
1 / 5 (2) May 15, 2014
ERROR Not only in this. In Others Too!
/other/2014/4120fusion_crushed_usadothtml on larouchepubdotcom says:
"Nuclear fusion has been at Mankind's Fingertips for Decades; Could Already be providing the World with Virtually Unlimited Energy; it has been deliberately suppressed under a top-down Imperial policy"

To Appear In May 16, 2014 Executive Intelligence Review Issue by Megan Beets
THE SUPPRESSION OF FUSION: How the Malthusians Crushed The U.S. Fusion Program
mohammadshafiq_khan_1
May 15, 2014
This comment has been removed by a moderator.
swordsman
not rated yet May 19, 2014
"Tipped" is mostly correct. Background noise is generally the main contributor to S/N ratio. However, there may be many other noise sources, depending on the environment, such as local or distance noise sources. You also have to take into consideration the character of the signals that are being measured. For instance, are you measuring voltage or power? That makes considerable difference.
vidyunmaya
1 / 5 (2) May 20, 2014
OM COSMIC DIVINE FIELD PRINCIPLES- BEYOND MILKYWAY- BY VIDYARDHI NANDURI
http://www.youtub...e=relmfu
Need to introduce prime concepts through base concepts