Seminar / Webcast / My Twitter
Slides available
Go to the Seminar page for the slides that there presented today.
The liveblog
(For details about the seminar, see below. Some links to photos on the twitter feed.)
The most recent updates are at the top of the page. All times are CERN times.
More updates will be posted as they arrive. Thanks for reading!
Questions
15:49: Rolf says “No more questions, so final remarks”. Great to have first results, remember they are preliminary and with small numbers. “Keep in mind we’re also running next year.” “The window is getting smaller and smaller, but it’s still alive!” “Stayed tuned for next year”.
(G) means that Guido answered, (F) means that Fabiola answered. Thanks to Rozmin Daya for providing more detailed transcripts of the questions and answers. (Questions are ordered so that the most recent question is at the top of the post.)
Q: Regarding the crystal calorimeter, by how much has this improved your endcap H->gamma gamma mass resolution? Did I understand that you understand the scale to 0.1%?
A(G): No. for the scale, if you take a look at the plot on slide 33, this rms is what you’re able to acheive. In the endcap there’s a lot of progress, the scale does not evolve in big jumps. We have lots of room to improve. Up to now we were limited by this phenomena related to transparencies. Our tracker is fantastic but introduces extra material. Have to understand material and conversions more. We did this for barrel but must do more aggresively for barrel.
Q: For ATLAS, you have exclusion at 115.5 GeV. Is there a way to have some kind of Look Elsewhere effect for negative fluctuations? For CMS, you have excess everywhere between 115-135 GeV. Can it be that you’re simply misunderstimating background?
A(G): it could be, but we should be really precise. I cannot exclude but I’ll give small probability to this. Atlas: as far as I know we don’t have a LEE for this exclusion, but it’s an exclusion at the 95% confidence level.
Q: In the 4l final state analysis, how much would you lose if we count only 2/3 events (to ATLAS)
A(F): It would go down to 1.5sigma to 1.6sigma
Note: This question was motivated by fact that now there is discrepancy between ATLAS and CMS that wasn’t there at HCP time.
Q: What is the ultimate scale energy energy scale precision in the gamma gamma, because you were showing 0.5GeV from the Z.
A(F): Uncertainty on photon energy scale is more. It’s a few parts per million on the Z peak, but it’s below 1%. When we transfer to the photon, we have to take into account that we use the Monte Carlo simulation. We vary the material in the simulation, and we end up with a few parts per million to 1%.
Q: it’s interesting to understand what are the signal resolution for the few events in the 4 lepton final state.
A(F): The mass resolution is typically about 1.9 GeV for muons, and 1.7GeV for electrons.
Q: Question for CMS. Did you try to extract the 90% exclusion limit for gamma gamma? It’s a bit close to the ATLAS excess. Don’t understand the strategy for W->lnulnu.
A(G): We use continuous approach: cut based, and then the invariant mass of two leptons in a boosted decision tree.
Q: I’d like to understand looking at gam gam fit and the use of exponenetials to describe the background. Choice in CMS was not to do that. How confident are you to do that, knowing that QCD bg not well modeled by exponential?
A(F): You’ll find the slide in the spares. We tried several functions. We tried using background coming from Monte Carlo simulation, and also adopted a conservative estimation on background by taking as background uncertainty in a bin of 4GeV the difference between exponential and the expectation from Monte Carlo generator. We get consistent results using other functions.
Talks
15:36: Finish and applause. Rolf gives overview and opens up the floor to questions.
15:34: Two excesses at 119.5GeV and 124GeV. Both excesses seem compatible with a Standard Model Higgs. 95% confident limits include 127-600GeV. Some excess is present in all 5 channels. Statistical significance of 2.6sigma locally and 1.9sigma with Look Elsewhere Effect taken into account.
15:30: Modes split by resolution. Low and high resolution channels agree that something is happening around 125GeV. Maximum local significance is 2.6sigma. With look elsewhere effect it’s 1.9sigma in low mass region. Expect 2-3 sigma effect in region 115-127GeV. Look for best fit of Higgs cross section, shows best agreement at 124GeV.
15:29: CMS more sensitive than Tevatron experiments combined! Expected exclusion is 117-543GeV, observed is 127-600GeV. What is stopping CMS getting lower limit? There’s some kind of bump there in the low mass region…
15:26: First glance at invariant mass plot. Exclusion plot looks like it shows excess at 125GeV, but deficiency at around 128GeV. Interesting, given what ATLAS saw!
15:25: H->gamma gamma analysis. Improvements in the vertex identification, energy resolution. Vertex finding efficiency gives roughly 80% or better in all data periods. Resolution measured using Z, W decays and pi0 decays. Laser signals used to correct for transparency measurements. A lot of work has gone into understanding these issues!
15:20: Putting limits using H->ZZ* mode, one of the most powerful modes. Expected to exclude the ranges 130-160GeV, 182-420GeV, observed exclusion in 134-156GeV, 180-395GeV and 340 460GeV.
15:19: H->ZZ*(llll) and H->gammagamma have excellent resolution. H->ZZ* is the “golden channel”. Expect 67 events, observer 72 events in full mass range. In the low mass region (mH<160GeV) CMS has observed 13 events, expected 9.5 events.
15:13: H->bb mode. Very challenging! Huge background from QCD processes. Look for associated production of a boson with the Higgs boson. Better sensitivity, but lower efficiency. Require a very boosted W or Z produced in association. (pT of 100-160GeV) 5 sub channels of H->bb with associated production.
15:12: H->tautau mode. Slight hints of excess. Limit plot shows gentle excess across the low mass region (110-150GeV) compared to expectation.
15:08: Using multivariate analyses for the H->WW* state. Cut and count analysis shows most backgrounds removed by a handful of cuts. (Standard Model WW production dominates to the end. Expected exclusion is 129-236GeV, observed is 132-238GeV. Then using a boosted decision tree, split samples into different topologies. Look for discrepancies in the BDT spectrum Expected exclusion is 127-270GeV, observed is 129,-270GeV. Looks like a small excess just below 127GeV!
15:06: Now onto H->WW* analysis. Large non-resonant background from Standard Model WW processes. Angle between leptons can be used as a discriminant. Leptons emitted in small angle, so invariant mass of leptons not very large (it's all about spins of boson!) MET can be used to discriminate against background.
15:03: Monte Carlo simulation plots shown for events. Topological constraints useful for removing background. H->ZZ(ll,qq) mode extended to low mass region. Study at high mass includes H->ZZ(ll,tautau). 10.2 expected background events, 10 observed, so not sensitive in this mode yet.
14:58: 8 independent decay channels modes shown in a big table, with their sensitivities. 4.6-4.7fb^-1 of luminosity used for each of the 5 main modes (H->gamma gamma, ZZ*, WW*, tautau, bb) Resolution is 1-3% for gamma gamma and ZZ* final states. All 8 analyses made it to preliminary results to be shown today.
14:56: More than 90% data taking efficiency in each mode, and 91% overall. Impressive! Analysis requires good understanding of backgrounds and object reconstruction. Good agreement with data for identification efficiencies up to hundreds of GeV. Standard Model cross section plot shown. CMS agrees with data across all the processes, with a slight deficiency in ZZ production.
14:50: Guide starts, outlining the CMS collaboration and the detector. Overview of the Standard Model Higgs boson. Showing results up to 600GeV. Different production modes give different sensitivity.
14:52: Flashback to slides from a year ago, showing expectations. Expected sensitivity down to Standard Model across the whole range when combining channels. Projected significance decreases sharply in the low mass range. Sensitivity will come from combining channels.
14:48: Finish and applause! Guido takes the microphone. And goes through Fabiola's slides by mistake!
14:45 With current data set ATLAS has excluded 112.7-115.5GeV and 131-453GeV (except for 237-251GeV) ATLAS is now competing with LEP's low mass results! There is a large deviation in p0-values at 126GeV. 1.9e-4, or an excess of 3.6sigma (gamma gamma 2.8 has sigma, ZZ* has 2.1, WW* has 1.4sigma)
Updating all other analyses for full data set. We need more data in 2012 in order to confirm if this is the Higgs. 126GeV is a nice mass for the Higgs- it can be probed with lots of modes (gamma gamma, ZZ*, WW*, bb, tautau).
14:40: Apologies, connectivity issues.
Now discussing H->ZZ* analysis. Statistics limited background studies for SM ZZ processes. Electron identification efficiency comes from J/psi, W and Z decays. Covers wide range of transvere energy (up to 50GeV). Monte Carlo simulation tracks particle identification well as pileup increases- we understand the detector very well. Isolated muons selected, isolation performance looks impressive, even as pileup increases.
Simultation gives mass resolution of about 2GeV, 85% of signal falls within two standard deviations of mass point. 71 events seen in the full range, expected background is 62 events. In the low mass region (gamma gamma at 126GeV is 2.8sigma! If it's due to background only, it's a very large fluctuation. There are nine categories of photon, with the background modeled with an exponential function, and Crystal ball+Gaussian for signal. Excess shown at 126GeV
14:27: Discussion of the angle measurement. Need to know position performance in the calorimeter. Resolution of position of primary vertex is ~1.5cm. Potentially large background from jets and hadrons. The faking is rare, but the rate of production of jets is orders of magnitude larger than the rate of Higgs boson production.
14:26: Sensitive at lower energies. Different from previous channel, need good resolution of photon measurements. Irreducible background from Standard Model gamma gamma, also some fake gammas from jets. Mass resolution and positive robust against pileup. About 5GeV width in the invariant gamma gamma mass (in simulation, based on knowledge of detector.) Energy scale known to 0.5%, about 1% for linearity and uniformity. Z->ee mass shown, good performance there. Knowledge of how electrons interact inform energy scale for photons.
14:19: Discussion of Standard Model backgrounds for WW* channel. Turn on of ttbar background for this mode at Missing Energy (MET)>50GeV. MET strongly affected by pileup.
Expected background: 76 events, Data seen: 94 events, Expected signal :19 events. Cannot improve limit with this mode alone.
14:17: There are lots of backgrounds to consider! Concentrating on the gamma gamma, WW* and ZZ* modes. Backgrounds are jets, photons and W/Z bosons.
14:16: Huge efforts go into understanding the detector. As the regions of the Higgs search change, the requirements of the analyses change.
14:12: Outline of Higgs search motivation. The two photon sample is most sensitive at low mass ranges. Massive vector bosons sensitive at higher masses. Theorists have been working hard to update their expectations. The allowed region is small. We'll make it even more smaller today, and maybe see something very interesting in there as we do!
14:10: "The Standard Model works at 7TeV. Very Good." Good performance of Standard Model processes. We must understand these to understand the backgrounds, and also to calbirate measurements.
14:08: Discussion of pileup, the price we pay for high luminosity. We increase the number of events we record at once by having several interactions per beam crossing. A big challenge at working at the LHC, and a challenge we meet. Triggers are closely monitored to pick out the most interesting results.
14:06: Understanding of the search is "well advanced". Fabiola expresses thanks to the LHC team. Data taking efficiency is 93.5%. Good quality data fraction is greater than 90% for all analyses.
14:05: First two slides. Fabiola explains the importance and difficult of Higgs searches. The first slide shows plots from several different analyses with data and Monte Carlo simulation.
14:01: Rolf introducing Fabiola and Guido. Huge round of applause for all the experts and LHC team. Building up the moment with a great sense of community. In spite of the competition between ATLAS and CMS, we're here together to present and see the results together.
Before the talks
13:59: One minute to go. Both talks and both speakers ready.
13:54: Just spotted Guido Tonelli, the Spokesperson for CMS and the second speaker today! Both he and Fabiola are looking smart, and ready to give us the facts.
13:45: Fabiola's talk has been copied and it is ready for her. She taking a sip of water and chatting with Rolf and technical support.
13:41: Experts from the LHC are here too. They have worked very hard to make sure the machine works for us, and we've had fantastic running this year. We must not forget their role in this work!
13:40: Fabiola Gianotti, the ATLAS Spokesperson is here. She will give the first talk in about 20 minute's time. She's chatting to the Director General of CERN, Rolf Heuer. She's smiling, but if I was in her position I'd be quite nervous right now! She's taking off her coat, looking at the microphone and so on. She's given many talks before, so she knows what's she's doing. Still, it must be nerve-wracking for her!
13:32: The webcast is now available! http://webcast.web.cern.ch/webcast/
13:30: A “delegation” of very smartly dressed people are arriving in the front rows! They’re more smartly dressed than most physicists, and they have reserved seats, so they are probably management, dignitaries etc.
13:27: Of course big names from CMS are here as well, including Albert de Roeck, Jim Virdee, and Gigi Rolandi! (Being an ATLAS member, it’s easier to recognize other ATLAS members!)
13:22: ATLAS Higgs Group Conveners, including Bill Murray and Eilam Gross, arriving now. These are the people in charge of the various Higgs searches at ATLAS. Some of the analysts are here as well. Lots of big names arriving.
13:14: Some thoughts about the media, science, and what we can expect to see today. The physicist sitting next to me asked about my blog and twitter feed, and we started discussing how pressure from the media can affect what scientists do. While it’s true that we love the media to be informed, we don’t change our results or interpretation based on public opinion. The results we see today are going to be exciting, but we need to be careful and do a proper job. If it’s not 5σ yet, it’s not a discovery yet. We’ve been searching for the Higgs boson for decades, so we want to get it right and we don’t want to sacrifice our standards for the sake of getting in a few months early. If we see the Higgs in the summer (and we probably will, if the rumors are to be believed) it will be the most important discovery in high energy physics since the W was discovered in 1983!
(If we don’t see the Higgs, that will also be an important observation, as it will tell us there is something else out there . Convincing bumps have been known to disappear when we add more data.)
12:53: The projector is being tested. As soon as the webcast is available I’ll update to let you know.
12:51: Katie snapped a photo of me and Pauline! It’s on yfrog.
12:28: I’ve just heard that the security staff at the door are no longer allowing more people to enter.
12:27: This seminar is one that should not be missed. Looking through the audience I’m glad to see most of my closest colleagues have found somewhere to sit, including @marktibbetts, Matt, Tina, Rozmin, Catrin and Sudan. We are normal physicists, fighting for whatever seats we can get. The first three rows have been reserved for special guests, representatives etc. There are people sitting on all the stairs. We’ve all heard the rumors. None of us has heard the official results form both experiments yet.
11:57: There are lot of people from the blogging community here, including Pauline Gagnon (to my right), Anna Phan (to my left) and Seth Zenz (in front of me)! You can follow Seth’s tweets, @sethzenz
11:42: Seating is at a premium, I just someone bring their own chair in!
11:31: I’ve been here for an hour now, and the auditorium is nearly full. There are roughly 10 seats left (none of them have desks, or power supplies.) Nearly everyone here has their laptop with them, it’s like a commercial for Apple! Staff are checking IDs at the door. There is a lot of chatter here.
Details about the seminar
Today sees CERN’s seminar on the update of the Higgs search. I’ll be updating this page as the information comes in. Refresh this page to get the updates! The most important points will be also be tweeted.
Time
The seminar will begin at 14:00 CERN time. (08:00 East Coast, 05:00 Pacific)
There will be a talk from the ATLAS spokesperson 14:00-14:40 giving updates on the ATLAS search, and then a talk from the CMS spokesperson 14:40-15:20. There will then be questions and answers to both speakers until 16:00.
Links
Here are the links to:
Seminar page
Webcast video
Follow the updates on twitter using the hashtag #higgsliveblog (my account there is @aidanatcern.)
The seminar page also has a chat room!
Shortlink for this page: http://bit.ly/u0wALv
Shortlink for the seminar page: http://bit.ly/s5X4Zm
Shortlink for the webcast: http://bit.ly/q2QB
Connectivity
We’re expecting a lot of internet traffic at CERN today, so there is a small possibility the network may get jammed for a few seconds from time to time. Thank you to Kevin for allowing caching of this page so that it can still be accessible in case of any problems.
Please report any errors in the transcript of this blog post in the comments.
Aidan Randle-Conde
