Quantum Diaries

Have you been paying attention to the LHC? Sure, you’ve been thinking about the scientific results being derived from last year’s data. And you are looking a few months down the road to the upcoming major conferences, now only a little more than two months away, when we might get some interesting news. But right now, the LHC has been running, and running well. Consider this: here is a plot of the integrated luminosity as a function of time for 2011:

And here is the same for 2012:

It is important to note the “preliminary” on this plot — all experiments are working to verify their luminosity calibration. But one can see that the integrated luminosity for this year at the end of April 2012, about an inverse femtobarn is about what it was for last year in the middle of June 2011. In all of 2011, we recorded “only” about five inverse femtobarns. (Dear LHC: could someone produce a plot with the integrated luminosity for both years on the same set of axes? Then I could make this point more easily.) We are recording data at a much faster pace than last year, and that can go straight into the physics bottom line. From what I have heard, the operating conditions of the LHC have been particularly good — the vacuum inside the beam pipe has been better than expected, which means that it will be easier than anticipated to increase the beam currents, and thus to increase the instantaneous luminosity more quickly.

This is important, because we’re about to hit the important big sprint of the year. The LHC has been doing machine studies and a technical stop during the past week. Regular operations for physics will restart around Monday. To be in a position to firmly observe a Higgs boson (if it exists) this year, we need to accumulate 6-7 inverse femtobarns by the “first breakpoint” in late June, under two months away. That is, we will need to accumulate as much data in the next two months as we did in all of last year’s run.

Can the LHC do it? Based on what we’ve seen so far this spring, I think we can try to be optimistic!

As has been reported elsewhere, the LHC is off and running again. Yesterday we saw the first stable beam collisions of the year. So far, collision rates are extremely small, and the detectors are just being roused from their winter slumber, so there is certainly no physics news to report yet. Over the next few days and weeks, the LHC will, fill by fill, increase the number of proton bunches circulating in the machine, and thus the collision rates. Meanwhile, the experiments will check that all of the detector elements are functioning and calibrated, which will allow us to get back to our full menu of work.

So what can we expect in the year to come? Here are a few things that I could think of.

OK, readers — what are you expecting from the LHC this year? We welcome your comments.

At the start of each calendar year, the CERN management holds a workshop in Chamonix to discuss the LHC run plan for the coming year and beyond. This year’s meeting was held two weeks ago, and this past week CERN announced the outcomes. Now, after last year’s Chamonix, the plan came out differently than many of us had been expecting. But this year’s workshop results were consistent with this year’s rumors.

There is a clear physics goal for this year: both CMS and ATLAS should each individually be in a position to discover the standard-model Higgs boson, if it exists. There are two ways that the LHC will try to make this possible. The first is to deliver as many collisions (i.e. as much integrated luminosity) as the LHC can manage. The integrated luminosity target for this year is fifteen inverse femtobarns for each experiment, three times as much as was delivered last year. It will still be a challenge to discover the Higgs with that much data; the experiments will have to run efficiently and the experiments will have to be as clever as ever. But it is possible. CERN is also prepared to extend the LHC run if necessary to meet this luminosity target. This is important, as the LHC will enter a long shutdown after 2012, so this year is our last shot for a while at making a discovery, of a Higgs or anything else. We should remember that last year’s target was a mere one inverse femtobarn, yet we got five times that. Can we hope that the LHC will outperform expectations again this year?

The second way to improve our chances of discovery is to raise the energy of the beams. The production rate for the Higgs and many other hypothetical particles increases with beam energy. Thus the LHC will run with 4 TeV per beam rather than the 3.5 TeV of last year. The operational experience of the past two years gives the LHC physicists confidence that this beam energy will be safe for the machine. This means that the LHC will probably never run at 3.5 TeV/beam again; the data we have recorded will now be unique in human history. It means that we’ll have to think about how to juggle resources so that people can look at both the old and the new data, and how to properly archive it for future use. Also, all sorts of measurements that we have made before at the LHC become new again: we can ask how does the production rate for phenomenon X change as you change the beam energy from 3.5 TeV to 4 TeV.

One change that the experiments had hoped for, but will not come to pass, was a change in the time interval between collisions. It was 50 ns during 2011, and it will stay that way. That means that we are now expecting an average of 30 simultaneous proton-proton collisions per beam crossing. Had the bunch spacing been reduced to 25 ns, we could have hoped to record a similar amount of data, but with much simpler events. However, the LHC experts weren’t sure that they could provide as much integrated luminosity at 25 ns spacing as at 50 ns; it is a very different way to operate the machine. Integrating data is the need for the year, so 50 ns it is. The experiments have shown that they can handle the complex events, although it would be a stretch to call it a pleasant experience.

Finally, the plan for the longer term was sketched out. The LHC will enter a “long technical stop” (as CERN likes to put it) at the end of the year, which will go on for twenty months. Given that we’ll need some time afterwards to re-commission the accelerator and the detectors, it’s probably two years from “physics to physics.” This will give the machine and the experiments time to implement some needed and useful upgrades and repairs. On the machine side, this includes the preparations to run the LHC at much closer to the design energy. That is 7 TeV per beam, although it is sounding like 6.5 TeV/beam is much more likely to be the safe operating point. At this point, we can only guess what the particle physics landscape will look like, but a higher-energy LHC will allow us to explore it thoroughly.

That’s the plan — let’s get ready to re-start the search for new physics in under two months!

Are you addicted to YouTube? No, I wouldn’t say that about myself, but gosh, it’s rather amazing what you can find on there. At home with the kids lately, we’ve been looking at classic bits of The Electric Company, the 1970′s Children’s Television Workshop educational show which spans the period of late Tom Lehrer to early Morgan Freeman. Part of what makes YouTube great is that it’s so easy to use. You put a phrase into the search window, and some computer somewhere (don’t ask me where) quickly finds the data that you are looking for. Then you just click a button and the videos come streaming onto your computer, without a whole lot of effort from you. You don’t have to know what computer disk the file resides on, or the directory structure of that computer. For all you know, the video might be coming from several different computers at once, with the source being adjusted in real time to give the best streaming performance.

Now, compare that to how we go about getting our data in particle physics experiments. Back in the day, you definitely had to know the exact directory and exact file names of the dataset that you wanted to analyze, and then carefully type that into your computer programs. A single typo could destroy hours or days of computing effort. We’ve largely gotten past that — we have better technology for file catalogues, such that you can just specify the name of a dataset, and all the file names will be looked up for you. But we are still largely constrained by “data locality,” the requirement that your analysis program must be running on a computer in the same room as the computer that has the disk with your data on it. This constraint leads to a variety of optimization problems. What if a dataset gets popular all of a sudden — are there enough processing resources in the right place to handle the demand? Can you get more copies out to the bigger processing centers quickly? Are you then under-using other centers and letting CPU cycles go idle? If you want to run on a given dataset, you might know which computing sites have that data, but how do you know which has the most available resources right now? And finally, what if data at a site gets corrupted? Will all the jobs running in that computer room start failing? Needless to say this doesn’t sound like YouTube at all.

I and some colleagues are working on a project that tries to change this. We’ve called it “Any Data, Anytime, Anywhere,” as our goal is to make it as easy to access LHC data as it is to access a YouTube video. At the heart of the system is a “redirector,” a system that serves as a giant index of files that reside at computing sites all over the country. A computer program asks the redirector for a file, the redirector finds an optimal source for the file, and the program then reads the file from that source, without the user having to know where the file actually is. That means that the source could be thousands of miles away, and the only way for the remote reading to be efficient is for it to be nearly as fast as reading from a computer in the same room, so some effort has gone into making that happen. Once you have removed the data locality requirement, all sorts of things are possible. If a file is corrupt at one site, it could introduce a fallback mechanism so that a read failure results in an attempt to get the same file through the redirector instead. If a particular site gets overloaded with jobs, we could start to migrate them to a less busy site, even if that site doesn’t actually have the data that the jobs want; they can be obtained through the redirector instead. That could lead to a better global balancing of supply and demand for resources. While we imagine that it’s computers at CMS institutions that will be reading the data, there’s nothing to stop any computer anywhere from reading the data, even if it is not part of CMS. That could really fulfill the promise of grid computing — if we can borrow a computer for a few hours, we can use it to analyze CMS data even if that computer starts out knowing nothing about CMS. It also gives us a straightforward way to use cloud-computing resources, if that were to turn out to be cost effective.

And on top of all that, what stops this from being limited to the LHC? Many disciplines have large datasets that need to be analyzed by distributed teams of scientists. In principle, they could use the same infrastructure. We’re hoping that this technology could eventually be used across the sciences and even into emerging fields like digital humanities. If that were to happen, then researchers from all sorts of disciplines could consider themselves Easy Readers, at least as far as their data is concerned.

Lincoln, Nebraska, where I live, is on the western end of the Central time zone, and as a result, the sun goes down pretty late on the clock here. Even at this time of year, sundown isn’t until 5 PM, and it’s not really dark until at least six. We usually get home with the kids around five, and then we do dinner and playtime inside until bed. That means that the children, who are five and three, are rarely outside when it is really dark out, and they don’t get to see the stars, beyond the bright planets, very often.

The past weekend was an exception; it was Chanukah and there were many evening celebrations, as you are supposed to light the candles at sundown, so we were out past bedtime. On Friday night, as we went out to our car to drive home, my daughter, the older kid, looked up at the cloudless sky and marveled at the number of stars that she could see. I looked up too, and took the opportunity to point out one of the few constellations that I can identify, Orion. (Whenever I think about Orion, I think about John Guare’s “The House of Blue Leaves” — sorry.) “See, it looks like a person, with a top part and a bottom part, and those three stars are a belt,” I explained. My daughter looked at this a little more, and then asked, “Does the world want it to be like that?”

Interesting question — what she meant was whether the stars were intentionally arranged in the shape of a person, or whether it was just something that people made up when they looked at the stars. The answer is the latter, of course, although perhaps the ancients thought differently. Our conversation for the evening went on to other topics in astronomy (“Planets are round,” she said, “so it’s very hard to stand on them.”), but I kept thinking about what she had asked me.

As scientists, we collect data from the world around us, and try to make patterns out of it that we can understand. These patterns are theories, really, and as more data come in, we re-evaluate the theories to see if they are still consistent with the data. Do all the stars make shapes that look like familiar things? Are all of the measurements from the LHC consistent with a Higgs boson at 125 GeV? Are we humans just imposing an anthropic view onto the world? Measurements throughout particle physics, not just at the LHC, seem to support the idea of the Higgs mechanism. Is that consistency just a pattern that we have invented? Or does the world actually want it to be like that?

A year from now, we hope to have an answer to this question. As we head into 2012, a potentially decisive year for particle physics, I hope that all of our Quantum Diaries readers have the opportunity to ask, and answer, their own questions about what the world wants it to be like.

(With apologies to Michael Turner….)

Here, some rapid post-game analysis for all of you. First, a big thanks to fellow US LHC bloggers Aidan Randle-Conde and Seth Zenz for their tweeting during today’s seminar, which made it easier for all of us in Nebraska to follow what was going on. Based on the social media, press releases and so forth that I’m following, this is a very big day for the field of particle physics, and it’s fun to be a part of it.

Let’s remember that the LHC only ended the 2011 proton run about six weeks ago, and in that short time since CMS and ATLAS have analyzed all the data recorded. You almost never see a turnaround as fast as that, given the data processing required and the careful validation that needs to be done of the data and then the analyses themselves. Congratulations are very much in order to the computing teams for the experiments, and all the people who are checking data quality, and all the people who stayed up late making the plots that were shown in today’s presentations.

The boilerplate summary has already been said, but I haven’t had the opportunity to say it yet: today’s results are certainly tantalizing, but it’s impossible to know what they will amount to in the long term. We’ve seen signals of this significance disappear before. Perhaps the more solid thing to talk about is the fact that the window of possibility for the standard-model Higgs is slowly but surely closing, as both experiments have now excluded a wide range of the possible Higgs masses. (The caveat here is the phrase “standard-model Higgs”; I noticed the other night that two teams of theorists — on the same day — posted articles saying that the addition of just one more particle to our theories could change all of these conclusions.) I’m not sure that when the LHC started up two years ago we would have imagined that we’d be able to make such strides so quickly.

In short, there is reason to be excited — but we don’t know what the reason is yet! We might be close to discovering a Higgs boson, or we might be close to excluding it. In either case, 2012 will be a decisive year for particle physics as we have understood it for the past thirty or forty years.

Now, on to the perhaps controversial part of the post. As I was trying to follow the talks today, I started to wonder — this experiment sees a peak here, this one sees one there. Is anyone being lucky? Face it, CMS and ATLAS got to record one set of data. If we were to record the same amount of integrated luminosity once more, we’d have a different set of events, and maybe we’d get some interesting events again, or maybe not. You can’t know. However, if we were to do the experiment again, we’d have the same detectors, and the same analysis techniques. The data are just some form of luck, one roll of the dice. The real figure of merit for how well the experiments are doing in the Higgs search is not the result you get from this one dataset, but how well you would expect to do for any given dataset of this size.

Fortunately the experiments tell you how well they expect to do — it’s encoded in the “expected limit” lines on the result plots. Here are those plots for the low-mass Higgs region of the search for CMS and ATLAS. Now, try not to look too hard at the black dots:

Just eyeballing things, I’d say that CMS expected an exclusion limit (in the absence of a Higgs) of 117 GeV, and ATLAS about 124 GeV. Obviously there are uncertainty bands on this…if we take the two standard deviation line from the expected limit and call that the worst-case scenario (or at least a worse-case scenario) then CMS would expect to exclude to about 133 GeV and ATLAS about 137 GeV. In the spirit of comity, I’ll declare this a tie. As an experimentalist, I would claim that the burning question for 2012 is not what the mass of the Higgs boson is, or whether it exists at all. Instead, we should be asking how quickly the experiments can push that “expected” line down such that there is the potential of excluding the Higgs with the data in hand. It will be done with some combination of more data and more cleverness.

Maybe today we’ve been lucky enough to see the first hint of a Higgs boson. Or maybe not! My experience in life is that you do have to be a bit lucky to get ahead…but before you can be lucky, you have to be good. What we have seen today is that both experiments are downright excellent! Next year we’ll find out if we’re lucky, too.

While avoiding writing the final exam for my course (sorry students, I’m now almost done with it), I stumbled on this article in The New York Times about the problems of a deluge of data in genomics. At this point, genomes can be sequenced much more quickly than they can be analyzed. Indeed, the article reports that there is enough sequencing capacity in the world to fill a stack of DVD’s two miles high with data each year.

Sound familiar? (Including the DVD analogy?) Particle physics experiments face the same problem. At the LHC, we have particle collisions 600 million times per second. The four LHC experiments produce a petabyte of data (a million gigabytes) per second — if we were to keep every bit of data that the LHC produced. Obviously, we don’t do that; the data is heavily filtered by the experiments’ trigger systems, which reduce the data rate to 300-400 events per second per experiment. Now, that will still get you something like 15-25 PB of data per year, and a stack of DVD’s that’s several times higher than that of the DNA sequences. So we have the same problem, if not a bigger one — and that is after we’ve only kept one in a million collisions! Particle physics has long been on the forefront of data-intensive computing.

I’m no biologist, and I won’t claim that I know any more about genomics than I do about soccer or wide-area networking. But it seems natural to ask what (if anything) genomics can learn from particle physics in terms of data management. I could think of two ideas. First, must they really keep all the data? We throw away essentially every collision that happens (but for that one in a million), and can still learn a huge amount of physics. I think that this is to a large extent because we know what is interesting and what isn’t, and know how to throw away the boring stuff. For all I know, it might not be possible in genomics. The data you are throwing away might be the genomes of individual people, and if you really want to understand how one particular person works, you can’t do that. But if you are just looking for trends in the population, maybe you can.

Another idea is, can they make the data any smaller? In particle physics experiments, we do a lot of “zero suppression” up front, just throwing away the information from electronics channels that have nothing to say about a particular event before we even record the data to a disk. Then, when we process the data to estimate the energies and momenta of the particles produced in a given collision, we typically store even less information. The samples we present to analyzers are very compact, essentially down to the momentum vectors, and not carrying all of the channel-by-channel information about each particle. I’ve read that a lot of our DNA is actually “junk” with no impact on how biological traits are expressed. How much of this can be identified in a given genome and then safely be thrown away? Or, if you don’t quite feel safe about throwing it all away, could you just keep it in, say, 10% of the genomes as a safety measure? (For trigger aficionados, this would be a form of prescaling.)

I don’t know the answers to any of these questions, but perhaps the biologists would, or could at least use them to stimulate some new thinking. Just the other day, the boss was telling the intensity frontier workshop that particle physics is part of a fabric of sciences in which different fields make broad contributions to each other. Data-intensive computing could be considered one of the threads that holds that fabric together.

Thanks to Ruth Pordes, executive director of the Open Science Grid, for suggesting this as a blog topic.

It’s been reported already in other outlets (e.g. other blogs), but since no one has said it yet here, there will be seminars by CMS and ATLAS on December 13 at CERN on the status of their searches for the Higgs boson. Significantly more data than was shown in a combined result just two week ago will be presented. Stay tuned!

At long last, here it is! From the Hadron Collider Physics conference in Paris, and as documented by the CMS and ATLAS collaborations, the plot you have all been waiting for:

As we last saw, CMS and ATLAS had each set limits on the rate of production of standard-model Higgs bosons at the Lepton-Photon conference in August. Now, for the first time ever, the two collaborations have combined their results. Each experiment has recorded about the same amount of data, so to first approximation, this combination allows us to double the number of collisions that are analyzed, and thus to set more stringent limits on Higgs mass (or possibly to discover a Higgs).

Since one of my colleagues took me to task just this morning for these plots being impenetrable, let’s review what is being measured and what the plot shows. First, remember that pretty much everything in particle physics is a counting experiment. You record so much data, and then count the number of times you observe a given phenomenon in the data. On the basis of this, you can essentially say, “given that I’ve seen this happen X times, surely if I were to do this experiment over and over, it would be very unlikely for me to see this happen more than N > X times.” N is then the “upper limit” on the number of events that we would expect to observe. (I’m sure my statistical friends will forgive me for this hand-waving description). We can convert that upper limit on event counts into an upper limit on the cross section for the process; the cross section is essentially the probability for a process to occur, which is obtained after normalizing out how much data has been recorded. The vertical axis of this plot gives the upper limit on the cross section for Higgs production, normalized to the expected cross section that we calculate from the quantum mechanics of the standard model.

The points show the upper limits that are obtained as a function of putative Higgs mass. It is a different upper limit for each Higgs mass because as the mass changes, you have different Higgs production and decay rates and different sensitivity to those decays; depending on the Higgs mass, it can be easier or harder to observe. As can be observed, the points fall below y = 1 over a wide range of Higgs masses. This indicates that we are observing fewer putative Higgs events than we would expect from the standard model prediction, and thus we claim to “exclude” that prediction, and thus the possibility of a standard-model Higgs at those particular mass values.

One should also pay attention to the dotted line and the colored bands. The dotted line represents what limit we would be able to set if there were no Higgs boson at all, and all there were to observe were background processes that look similar to the Higgs but aren’t. The bands represent the one and two standard deviation uncertainties on that expected limit. In general, the limits set are about as good as those we expected to set. There are some excursions from expectations, but they are generally no worse than two standard deviations, which is not impossible. This gives us some confidence that the observed limits aren’t anything crazy.

By combining the results of the two experiments, a very wide range of possible Higgs masses is excluded. Neither experiment alone could produce a result this strong, and hence the great interest in the combined result, which took many months and much coordination between the two experiments. Each group of experimenters had to understand the others’ measurement in detail to be able to do the combination correctly. It’s a lot of work, but the improvement in the bottom-line result is worth it.

And what do we learn from this? It appears that if there is a standard-model Higgs boson, it must have a very large mass (which is disfavored by other measurements), or a mass between 114 and 141 GeV. Optimists will note that in that region, more candidate events are observed than would be expected from a no-Higgs scenario, although not with any statistical significance worth talking about. If one believes everything about a standard-model Higgs, then ATLAS and CMS are currently putting quite a squeeze on its properties.

Of course, that’s a big “if.” The contrarian in me likes to keep two things in mind. First, all of this statistical stuff is just a convention we’ve adopted to communicate with each other. (We’ll see if my statistical friends forgive me for saying that!) The definition of excluded is, in my opinion, rather arbitrary, and you could imagine doing things differently and coming up with a different range of excluded Higgs mass values. If we are are to claim a discovery of a Higgs boson someday, I would assert that the evidence will have to be even clearer than what can be obtained from such statistical analyses.

Second, why should we believe any of the predictions? They are cooked up from many ingredients, each of which have their own uncertainties with them. It is hard to believe that the predictions are to be wildly off, but how the physics really works might not be what the standard model says it is, and thus we have to keep an open mind.

Thus, even if we reach the point where we can exclude all reasonable values of the Higgs boson mass — a point that we might reach soon, given that the experiments have recorded at least twice as much data has have been included for this combined result — and we actually do exclude those values, the search will not be over! Even if the standard model is not correct and there is no Higgs as such, the signatures of Higgs production and decay are still interesting, and could still be an indication of some kind of new physics. Higgs or no Higgs, we have a very interesting few months ahead of us.