So the last hope collider people are getting desperate?
To explain a bit, the well known trouble with today's physical theories is, that on one hand we know that general relativity and quantum field theories do not fit well together, on the other hand we don't have a good idea of the way forward. So the hope was, that LHC would show something interesting, like supersymmetry, or at least deviations from the standard model. So far it only shows precisely what one expected before, so the theoretical problems are precisely the same as before.
Not really. We have several ideas going forward, it's just that no experiment so far has produced evidence for them. On the other hand, a null result is still a result. Of course it would be cooler to witness supersymmetric partners, but the fact that we searched and didn't find them at those energies is a result in itself.
Is is true though? I've been to a talk by one of LHCb researchers [1] where he said that they observe a pretty significant anomaly in meson decay (I believe [2] has more details) and expect to publish updated results later in 2018. Looks like there is some hope
null result is still a result. thats cute. but looking at the budget of LHC i would hope for a bit more actual result. instead of using backward logic to promote failure as success.
ah, i have this wonderful theory.... i didn't prove it yet. but that doesn't disprove it... >.> what happened to solving real-world problems, with practical solutions, and doing science along the way? we used to discover many interesting things about nature and the universe in just that way, and at the same time this attitude changed, the 'null result is also a result' cult started to eat our brains and pollute our thought.
> what happened to solving real-world problems, with practical solutions, and doing science along the way?
Imho, the most beautiful results originated from letting people just sit in a room and think, rather than putting pressure on them to generate something useful (e.g. publication pressure, shareholder pressure).
That's not a null result, but a negative result. A null result is a result that tells you nothing new (by definition). At least, according to the parlance in this thread.
Particle accelerator research has resulted in theories and equipment as a byproduct of what they found - either proving or disproving specific theories, or by producing new or unexpected results. The last of which results in new theories to explain those results. Those new theories producing “useful results”.
MRIs and PET scans were the result of theories that only came about as a result of data that came out of particle accelerators. The goal of the accelerators was never specifically “develop new medical imaging devices”, it was simply the “basic research” that makes thing like that possible in the future.
Promote failure? Do you even know what the scientific method is? They found the higgs boson! They proved some of the most difficult particles predicted by the standard model. How is that a failure????
Isnt it more so that LHC proved the Higgs-Boson but not much else and they were expecting to find quite a few things more at the energies they are running (at least according to the theories).
Sorry to be unaware, but what were/are they hoping or expecting to discover with these experiments, that is other than find new subatomic particles. I mean, they did find a whole set of subatomic particles and their respective spins, so no reason why they couldn't find something more. I know it's all for recreating the scene of the big bang or whatever theory that they hope to prove or disprove or figure out a new one.
I do know of the Standard Model and Beyond the Standard Model types but it really seems to boil down to discovering new theories perhaps. That could also be uncertain as a new model could ally with QM or not, again I don't know if the current Standard Model allies well with QM. I think it doesn't again ally well with GR which is the same problem with QM and GR. It's funny though how they all behave.
They predicted and found the Higgs, but things like more particles that helped explain why the mass of the Higgs isn't what it should be, hoped to find things like super symmetry, extra dimensions, strings, new combinations of old particles, but it literally found nothing. (Im not an expert but am paraphrasing talks Ive hear from Sean Carrol). All it found was the Higgs and nothing else, and I think that is more than puzzling for physicists. The good thing is particle and quantum physicists like puzzles.
String Theory is a theory in the same sense that Quantum Field Theory is a theory. You need concrete phenomenological models to make predictions. While there some ways to do that starting from string theory there is no one known model to rule them all, the way the earlier attempts like Georgie‘s SU(5) GUT Model was. For some flavors like F-theory it is even a research questions how to make certain predictions of matter content etc at all. This is made more uncomfortable by te fact that the most natural susy extensions of the standard model have been all but ruled out. This also eliminates one of the main reasons for susy: as a solution to the fine tuning problem.
Before they outlined what they were looking for. Now they're just sifting.
Most science doesn't allow you to propose "We will expose A to X and see what happens", but rather expects a hypothesis: "We expect Z following X>A because of [rationale]."
"Most science doesn't allow you to propose "We will expose A to X and see what happens", but rather expects a hypothesis: "We expect Z following X>A because of [rationale].""
Which I consider one of the major flaws of the current scientific practice. There's absolutely nothing wrong with trawling over data looking for interesting things. You just have to be more statistically careful. Huge amounts of science have been done by just trawling over things or rapid-fire throwing theories at the wall and seeing what sticks. Having to always call your shots means you can only slightly push the frontier back. There's a place for that, which is probably "the vast majority of experiments". But we need that more exploratory stuff too.
But somehow we got entrenched with a terrible oversimplification of science as "The Way Science Is Done", which hyperspecifies the parameters of the ways we can modify our confidence values in various theories. There's more valid choices than is currently considered valid, and we're missing out.
I wonder if this is due to the success of the current models (less hidden corners), academic incentives for paper production as opposed to breakthrough work, the ever growing specificity and cost of experiments (making less sure experiments unviable) or some combination of these factors.
This is a good question! It really depends how you're smashing the protons together.
Did you create a model and then smash them together and analyze them? So you are testing "I believe that if I smash protons at X energy level that they will cause y_0,...,y_n decays" This is standard hypothesis testing. You are looking in a specific place.
A brute force method in this area is more like "Well... I'm not really sure. Why don't we just smash these protons together at uhhh... idk X energy?" And then you get "well that was interesting" or "okay, that was expected".
This is overly simplified. Basically a brute force method is just doing and observing. Not really predicting things. You're searching a large solution space and so you just try random things and hope you get lucky. (they have some good guesses of areas too look, so it isn't completely in the dark)
It's two parts. The smashing is just gathering raw data. Lots and lots of data.
Then "brute force" approach has to do with what can be found in that enormous data set and how physicists have, until now, been looking for specific patterns from theories and expectations.
Basically, the LHC spits out data so your your Big Data looks small. The first pass was for data confirming/disconfirming some big theories where a specific peek at the data would give meaningful answers. So far the data shows that we had it right to start with.
Now, for the next pass, they're expanding how they look at the data to find anything 'unusual'. "Pattern recognition" in this case is a bit harder than just checking for a recurring value, or a "true" in the "Unified theory of everything?" field. We're a bit closer to what we would generally call AI, and at very least some interesting techniques with more interesting followup experiments to weed out false positives.
We have models which provide predictions about the signatures we would see in the detector if that model is true.
The "general search" doesn't target a particular model. The only model used is our Standard Model, treated as background. Traditional searches utilize one or more of many Beyond Standard Model (BSM) models to treat as potential signal.
It seems like an interesting problem. It has clearly defined requirements for success, lots of data with (one assumes) some hidden structure, and the structure is not yet known.
I've yet to read the article, but I'll hazard a guess. At the time of LHC's design and instantiation, the amount of raw data it generated was vastly more than could be stored and arbitrarily processed. So, it was quickly filtered and shunted into processes designed to look for certain things.
Communications and processing capabilities have increased substantially since then. One may now be able, with investment/expansion of those, to take a more "brute force" approach in mining closer to if not at the level of raw data and arbitrary -- or at least a broader range of -- associations.
It's not quite that simple. The CMS detector (one of the two main detectors/experiments operating at the LHC) spits out about 10TB of raw data per second. This is more than any system can handle at the moment. To deal with this overwhelming flood of data there are layers of "triggers" that effectively filter out the collisions that aren't of any interest. The lowest level triggers are actually embedded in FPGAs on the detector itself, then there is a high-level trigger that runs on an on-site server farm that does minimal event processing to see if the even is interesting.
All of this requires us to define what constituents an interesting event. The most basic trigger is called the "minimum bias trigger" because it requires us to make the fewest assumptions about what makes the event interesting. So even at the most general level a brute force approach will still have some bias because it will likely be using the minimum bias data. What makes it "brute force" is that they will no longer be looking for signals from specific new models, they will only be considering the Standard Model and essentially looking for outliers. The difficult part is that there is so much systematic uncertainty given the complexity of these detectors, which makes it hard to obtain statistical significance.
Thank you. I find your explanation enlightening and interesting.
I haven't spoken with him in some years, and I won't mention names, but a good friend worked on calibration of the CMS detectors (scintillation). The last time I did speak with him, that was a very interesting 2 or 3 hours. Although I don't recall us discussing the data processing aspects in any depth; my limited impression of those I mostly gained from random articles and whatnot.
These experiments typically run 9 months out of the year with (very roughly) 50% up time. That's (again, very roughly ) 25,000,000 seconds. That's 250000 petabytes a year [0]. And by "raw" I'm talking about the ADC signals from the detector hardware. For example, the readout from the 75 megapixel "camera" (i.e. the tracker) at the core of the detector that takes 40 million pictures a second. When doing analysis the collisions are reconstructed into "events" where all the ADC signals are combined to form particle tracks and energy deposits (allowing us to actually study the physics behind the high energy interactions). These reconstruction algorithms are very complex and statistical in nature so there is some bias associated with them, hence the reason the "raw" data is permanently saved along with the reconstructed data.
Compression is most definitely used for the data that is actually saved to disk but it is not done on the fly when trigger systems have to run in nanoseconds.
The data limitations are not really from a lack of foresight or proper budgeting. It was known decades ago how much data would be produced by the LHC, it just really is an insane amount of data.
[0] after applying triggers (i.e. filtering) only a few hundred petabytes of data is stored in a given year from the CMS detector.
It’s been ~10 years since I was at Fermilab working on the CMS Tier 1 team; when I was still new on the team, I expressed concern about data loss due to scheduled maintenance on our Nexsan clusters (coming from the private sector) and my boss would chuckle and say “Don’t worry, they’re always making more data”. Also, scheduled maintenance windows during the day! As an ops guy, I was spoiled for that period of time in my career.
I'm trying to understand your numbers, at 10TB/s and 40 million pictures per second, thats 250kB per picture. since there are 75 megapixel (or I assume 75 megasensors) how does that work for raw data?
Is the energy determined from calorimeters, or from curvature in magnetic field?
What exactly is the ADC measuring? deposited charge?
Since you are acquainted with the project, could you point me to a good pdf describing the detector and low level digitization architecture (pre FPGA) architecture?
You're reading a bit too much into the "raw data" part. An average event that is saved to disk is about 1MB in size but that is with compression. The uncompressed ADC counts obviously would add up to a lot more than that given the number of readouts in the tracker, which is only one of about a dozen sub-detector systems. So saving every event to disk, with compression, would amount to about 40TB/s. Since there are a number of caveats I was only using an order of magnitude to illustrate the scale of the datasets being handled at the LHC.
I'm sorry if I sound dismissive here, I don't mean to be it's just impossible to go into all the detail of these machines in an HN comment. They are unbelievably complex (and fascinating!). I encourage you to check out the CMS design doc that will cover most of it in extreme detail (it's almost 400 pages) [0]. There is also a much more general overview of the various detector subsystems that is considerably more concise [1].
To answer one of your questions though: the tracker is used to measure the track of charged particles through the magnetic field and determine their momentum. This isn't enough to tell us their energy though b/c we don't know how massive each particle is. This is why the tracker is surrounded by an electromagnetic calorimeter (which "catches" electrons and photons) that is surrounded by a hadronic calorimeter (which "catches" hadronic patricles like neutrons and protons). The information from these systems is combined to ID each particle and determine how much energy it had, whether or not it decayed, etc. This is generally how we study the physics involved in the collisions but in reality it is 100x more complicated than this and there are number of other detector subsystems involved.
I have the impression that the reason the data for a single event is so large is because it is ambiguous, i.e. the uncertainty of a possibly incorrect interpretation forces you to store a lot of analog values (so that the interesting events can be checked or re-interpreted later on), and this size is what forces you to store a lot of data, and this forces you to triage for interesting events.
Assuming this is correct, then if the uncertainty of the tracks and their energies could be decreased, the data would be much more compact (i.e. only store location, momentum, energy/mass of incoming and decay particles as opposed to all the possibly relevant for later analysis ADC values)
Assuming that this level of correctness could be achieved by solid-state track detectors [0], where the particles leave tracks throughout a solid material, would it not (in theory) make sense to 1) continuously pass a such a detector material above and below the collision point, 2) at high speed slice, 3) in massively parallel etch, 4) microscopically examine the accurate tracks (bubble chamber style) 5) digitally store them compactly by the above assumption that the event size in bytes would be much lower if we only needed to store few but exact parameters per event 6) remelt the preferably low melting point etched track detector material 7) feed the erased etched track detector back to the collision point in a continuous fashion.
Either the speed of the material should be quite high or the illumination zone or collision point engineered very compact. Different collision bunches would be disambiguated because the tracks point to a different origin point (since the material has moved in the 25 ns @ 40 MHz, so for sub micrometer collision point that is 40 meters per second! obviously the detector would benefit from being designed to steer the collision point so it alternates to different positions)
Obviously the biggest issue with this approach is needing a huge facility for parallel etching and inspecting the slices
the cooling time should not be an issue, since the circular buffer of remelted material could be made arbitrarily long
do you think it could make sense for a future detector to temporarily store the tracks physically in a solid-state nuclear track detector ?
It's an interesting idea but, if I understand it correctly, I don't think it would be able to capture the majority of particles emitted from the collision zone. Modern detectors are designed to be as hermetic as possible so they can capture as much of the transverse energy as possible. This is crucial because in the plane perpendicular to the beam axis (i.e. line of collision) the total momentum is zero. Thus we can use conservation of momentum to infer the presence of particles that the detectors struggle to find (like neutrinos). This, of course, only works if we're confident that we accounted for (almost) all of the particles coming out of the collision.
Here is a picture of a heavy-ion collision reconstructed from the CMS detectors to give you an idea of how much needs to be accounted for [0]. The light orange lines are the particle tracks in the tracker, the red boxes indicate the amount of energy deposited in the EM calorimeter, and the blue boxes indicate the amount of energy deposited in the hadronic calorimeter.
I should note that normal proton-proton collisions do not produce anywhere near this number of particles (what we call the event "multiplicty") but they can still produce a few hundred. Furthermore, there are, on average, 40 proton-proton collisions per bunch crossing (which is what dictates the 40 MHz rate). So every 25ns there is a bunch crossing and the detector "snaps an image". In that image there are typically about 40 different collisions. Correlating tracks to vertices is not that easy given these conditions and is further complicated by the fact that particles can decay mid flight, causing their track to suddenly change. It would be great if we could be absolutely confident about the physical quantities associated with all of the particles produced in a collision, that's all we would need to know and this is essentially what's produced in high-energy physics simulations. However, reality is a lot messier. In addition to what I previously mentioned, the interactions between the detector material and the particles themselves also creates a lot of noise in the system. All of these uncertainties make it infeasible to compress an event down to just the reconstructed physical quantities, there is too much uncertainty and I don't think it will ever be overcome.
Lastly, I should mention that the energy densities reached in the immediate vicsinity of the collision zones is high enough to destroy any material known to man. The center of heavy-ion collisions reach a temperature of over a trillion degrees. That's 100,000 times hotter than the center of the Sun. No instrument could ever be placed there to make an accurate reading. The best we can do is place detectors a decent distance away (IIRC it's about a few cm) and try to catch what comes out.
Those interested in a (somewhat) accessible view of where particle physics is at in a big sense could read this article from the head of our theory group at CERN.
It’s the machine. The machine has to be configured. Yes, there are group leaders who have the final say on the configuration, and no, they aren’t grad students.
But it's always been a brute-force approach which is usually used at CERN, for detecting and analyzing the particles that may arise out of the various particle accelerator collisions. It is however a wiser approach to probably involve the current hardware capabilities to crunch the vast amounts of data through some statistical analysis or algorithms to sift through so much data without much human intervention and obtain results that have lesser false positives, as those are one of the drawbacks of using the brute force approach. Further research in these areas which could then help improve the accuracy of these models an then be reinforced into the chain and obtain only the required data to then manually observe.
The article never outlines what the alluded-to downsides are (p-hacking, spurious correlations) which makes me wonder if the article’s sourcing leaned too heavily on advocates of this theory-free approach, which was the subject of a Wired cover story a few years ago, so it must be true, amirite?
Reading the article makes me wonder what Woit (author of Not Even Wrong book and blog) thinks about this and how it dovetails with the epistemological morass the string theory people have gotten themselves into.
A somewhat early harbinger of the U.S.'s decline in public funding of fundamental science -- and all the practical discoveries, engineering, and knowledge that derive from same.
From random news stories I encounter, it appears that "they" keep trying to shut down FermiLab. However, despite being "older" and "smaller", FermiLab continues to make important contributions. And among other things, they contributed significantly to the Higgs Boson work.
Not mentioned in the Wikipedia article, was the debate over whether the SSC should even have been located in Texas -- apparently, there were significant political factors in the decision. IIRC, there was counter-argument that it should be located as close as possible to FermiLab, for synergy and efficiency in continuing its work.
There are the cited "costs". However, these project seem to invariably produce a lot more practical value than they consume. Cryogenics, superconducting technologies, data processing and communication, and on.
As I've said before, if you appreciate your medical MRI...
Not to mention NMR contributions to chemistry and biology...
Etc. It all hangs together.
And killing fundamental research is kind of like killing the goose that lays those golden eggs.
> apparently, there were significant political factors in the decision
As I recall the history this project is a near-perfect example of how pork-oriented-spending is not congruent with R&D projects. Everything from site selection to subcontracting was divvied up to appeal to the relevant congressmen. These unnatural constraints caused delays, logistical issues, and massive cost overruns. Their ability to solve those issues was hamstrung by the projects organization. Thanks to the dependencies they were a ways in before this all blew up and the project had to be stopped.
Building something like the SSC or LHC is fundamentally hard. Making someone do something that is hard while twirling a baton for political pleasure, much less getting critical components from two suppliers who can't integrate until construction, almost guarantees failure.
The correct approach, ironically, is perfectly embodied by the post WW2 machinations of the US government... Make the goals clear, dump appropriate money in engineering and technology organizations, let the nerds figure out the devlish details, and enjoy your A-bomb blowing up on time and schedule.
The article never clearly describes the brute-force approach. Is it just a matter of looking for statistical outliers along all the different dimensions they can think of? How would they then distinguish the "interesting" outliers from random flukes?
The gist is that you would roughly look for deviations from theory on one (probably small) dataset, you'd identify any regions of interest and then run a proper analysis of that region on a different dataset to see if the effect is real.
The problem with this approach that I see is that it's much more likely to pick up lingering detector issues than the regular test-a-theory. (The difference between looking for a specific thing in a specific place vs picking up anything unexpected.) I wouldn't worry do much about purely statistical artifacts because those can often be worked out with prescriptions and remeasuring. The systematic but not-understood biases are the ones that should plague this approach since in the extreme they would need an independent experiment. I wonder whether CMS and Atlas are sufficient for that.
To explain a bit, the well known trouble with today's physical theories is, that on one hand we know that general relativity and quantum field theories do not fit well together, on the other hand we don't have a good idea of the way forward. So the hope was, that LHC would show something interesting, like supersymmetry, or at least deviations from the standard model. So far it only shows precisely what one expected before, so the theoretical problems are precisely the same as before.