Lost quantum traces could reveal dark matter at the Large Hadron Collider

Particle physics is entering a new age of exploration. For decades, theory provided a map that pointed physicists towards new particles and told them roughly what to look for. Now, as the Large Hadron Collider (LHC) prepares for its most powerful incarnation yet, we have reached the limits of the map. In late June, the LHC shut down high-energy operations ahead of a major overhaul. When it switches back on in 2030 as the High-Luminosity LHC, it will produce around 10 times as many collisions as before. Somewhere in that sea of data, physicists hope, may be signs of particles and forces we have never seen. This situation leaves particle physicists with an unusual problem. The upgraded LHC will give them more territory to explore than ever before, but fewer clues about where the treasure is buried. Worse, its torrent of collisions will create an extraordinary computational challenge: any glimpse of new physics may be buried under billions of perfectly ordinary events. Advertisement New Scientist Live Sarah Alam Malik is just one of the fascinating speakers at this year’s festival, taking place in London from 10-12 October Sarah Alam Malik, a particle physicist at University College London, thinks quantum technologies could help. After more than a decade searching for dark matter at the LHC, she is developing quantum algorithms to spot unusual patterns in collision data. The hope is that, if physicists can record more of the strange quantum features inherent in particle collisions, her algorithms could probe those to search for new physics beyond the standard model – our best theory of the universe so far. New Scientist spoke with Malik about how quantum computers could help us parse this coming deluge of data for new particles and even reveal clues in the elusive search for dark matter. Thomas Lewton: Why are we sure that there are new particles out there to find? Sarah Alam Malik: Partly, because we can’t explain everything we’ve seen in the universe with just the standard model, the best theory of reality that we have so far. For example, we know there’s something that we’ve dubbed dark matter, which accounts for a quarter of the universe. It is some invisible scaffolding that holds together the universe at large, and we have evidence of it from the scales of individual galaxies – from the rotational velocities of stars within those galaxies – all the way through to the large-scale structure formation of the universe. So, there’s overwhelming evidence that there is something out there, some kind of unseen mass. One of the most compelling proposals for what it might be lies within the domain of particle physics, and is something we call weakly interacting massive particles (WIMPs). We think it might be some new kind of particle or an entire sector of particles that we haven’t observed so far, which may not interact with light, but have mass and interact gravitationally. A few theories have predicted the existence of WIMPs for different reasons. For example, a theory called supersymmetry predicted a whole host of new particles that helped explain why a particle we discovered in 2012, known as the Higgs Boson, is much lighter than what the standard model suggests. Among these new particles is the WIMP, which has all the right attributes to account for dark matter. The possibility that we could explain both dark matter and the Higgs’s light mass with WIMPs was called the “WIMP miracle”. Since 2010, we’ve been hunting for WIMPs at the LHC by smashing protons together and accounting for the energy carried away by the resulting particles. If some energy appears to be missing, it could mean an invisible particle – perhaps dark matter – escaped the detector. The LHC hasn’t found dark matter particles so far. But how are scientists planning to use quantum computers and AI to improve their chances? There’s now a huge desire to look for dark matter via a more bottom-up approach. Nature has just seemed far too clever for us so far. So, being more model-agnostic in our searches is, I think, the way to go. Instead of looking for particles predicted by our current theories, like the WIMPs of supersymmetry, you write down the simplest interaction that could take place between, say, the core constituents of protons and dark matter particles. Then you work your way up from there. Or you dispense with theory and let the data be your guide. There’s a lot of excitement about quantum computers in this regard. They might be one of the most disruptive technologies of the century, and we’re trying to leverage them to see if we can accelerate the process of looking for new particles in a model-agnostic way. One of the examples of such a technique is called anomaly detection, which finds patterns in information using machine-learning techniques. Your bank might use techniques like these to detect whether your bank card has been used fraudulently based on deviations from past behaviour, which they know very well. In my research, we are training algorithms on standard-model processes. The idea is that quantum computers may pick up anomalies that diverge from the standard model with greater sensitivity and accuracy. That sounds like a challenge. Tell me more about how it would work. We’re ultimately trying to draw some kind of “decision boundary” between two classes of data. This boundary helps us divide events at the LHC into two groups: those that are expected from the standard model, and potential anomalies or deviations from that. We’re developing quantum approaches that we think can yield an advantage. The key point is that the particle collisions at the LHC are inherently quantum processes – but when quantum particles pass through our classical detectors, they pretty much become classical data. A quantum system, unlike a classical one, can contain information about a range of possible outcomes and about correlations between particles. Once we measure it, however, we only record a particular outcome, and some of that underlying quantum information is lost. The hope is that by applying as little classical processing as possible to the detector data, we can retain more of the information preserved in the original measurements. In doing so, quantum computers and algorithms might be able to identify subtle patterns in this less-processed data and separate different kinds of events in ways that are difficult for classical methods. Exploratory studies in this direction look promising, but it is too early to gauge the extent to which quantum computing methods will be advantageous. People are also looking at quantum sensors and whether we can extract information from the system without fully destroying the delicate quantum processes. This is very much in its infancy, but it is an exciting possibility. Will these quantum algorithms be ready for when the LHC turns back on in 2030? Quantum computing is expected to really come of age in the next, say, 10 years or so. So, you would hope that there is going to be some confluence of those timelines where we are able to really leverage it to its full potential, once we have this enormous dataset to play with. Might we have got the wrong end of the stick entirely? What if dark matter isn’t a new particle at all? Maybe. One of the attractions of looking at dark matter is that it bridges the gap between cosmology and particle physics. Ultimately, that largest of scales is composed of these tiny particles that you’re studying at these particle colliders. If we discover something at a collider, and it seems to have the right attributes to explain dark matter, we still couldn’t immediately conclude that it is the substance responsible for the dark matter we infer from the cosmos. Collider experiments probe processes that unfold over very short timescales – around 10-23 seconds – whereas dark matter has shaped the universe over billions of years. So, a particle that looks promising in the lab would still have to be shown to behave in the right way over cosmological timescales. And it remains possible that dark matter isn’t a single new particle at all, but something more unexpected. Currently, there’s been this tension between two hypotheses: one is particle-based dark matter, and the other is modified gravity scenarios. These scenarios propose that the laws of gravity are actually different over large scales. Particle-based dark matter has been the leading contender so far, mainly because it can explain much more of the breadth of observation that we have, from individual galaxies and galaxy clusters to the cosmic microwave background radiation. You have recently written a book that takes a much longer-term view of physics. Has that changed your perspective on all this? My book A Brief History of the Universe (and our place in it) is about cosmic upheaval and our relationship with the universe: these big breakthroughs that overturned our worldview and also our perception of our place in it. So, it takes us from the ancient Babylonians, who looked at the stars and thought that they were gods or the messengers of great deities, through the ancient Greeks and the Islamic Golden Age and the Copernican revolution to the modern era. It’s made me realise that our frameworks have only ever been provisional, and [that we need to] treat our current framework as the same: that we shouldn’t get too wedded to our ideas; that we know they are liable to be swept away in the next scientific revolution; that they will be upended by future generations in favour of something that may be quite radically different from anything we can imagine right now, because they will have much greater knowledge and access to technologies that might light up the universe in ways that we cannot foresee. I really like a quote by astrophysicist Arthur Eddington, where he likens scientific discovery to putting together a great jigsaw puzzle. He said that one day you may ask a scientist how they’re getting along, and the scientist says, “Very well. I’ve nearly finished this piece of blue sky.” And the next day, you ask how the sky is progressing, and he says, “Oh, I’ve added a lot more, but it was sea, not sky. There’s a boat floating on top of it.” And next time, the boat might turn out to be a parasol upside down. A new discovery may not mean that the already arranged pieces have to be taken apart, but that the whole picture has to be reinterpreted. It’s just very exciting to speculate about how things might look thousands of years from now.