Computers Math

Physicists searching — unsuccessfully — for today’s most favored candidate for dark matter, the axion, have been looking in the wrong place, according to a new supercomputer simulation of how axions were produced shortly after the Big Bang 13.6 billion years ago.
Using new calculational techniques and one of the world’s largest computers, Benjamin Safdi, assistant professor of physics at the University of California, Berkeley; Malte Buschmann, a postdoctoral research associate at Princeton University; and colleagues at MIT and Lawrence Berkeley National Laboratory simulated the era when axions would have been produced, approximately a billionth of a billionth of a billionth of a second after the universe came into existence and after the epoch of cosmic inflation.
The simulation at Berkeley Lab’s National Research Scientific Computing Center (NERSC) found the axion’s mass to be more than twice as big as theorists and experimenters have thought: between 40 and 180 microelectron volts (micro-eV, or ?eV), or about one 10-billionth the mass of the electron. There are indications, Safdi said, that the mass is close to 65 ?eV. Since physicists began looking for the axion 40 years ago, estimates of the mass have ranged widely, from a few ?eV to 500 ?eV.
“We provide over a thousandfold improvement in the dynamic range of our axion simulations relative to prior work and clear up a 40-year old question regarding the axion mass and axion cosmology,” Safdi said.
The more definitive mass means that the most common type of experiment to detect these elusive particles — a microwave resonance chamber containing a strong magnetic field, in which scientists hope to snag the conversion of an axion into a faint electromagnetic wave — won’t be able to detect them, no matter how much the experiment is tweaked. The chamber would have to be smaller than a few centimeters on a side to detect the higher-frequency wave from a higher-mass axion, Safdi said, and that volume would be too small to capture enough axions for the signal to rise above the noise.
“Our work provides the most precise estimate to date of the axion mass and points to a specific range of masses that is not currently being explored in the laboratory,” he said. “I really do think it makes sense to focus experimental efforts on 40 to 180 ?eV axion masses, but there’s a lot of work gearing up to go after that mass range.”
One newer type of experiment, a plasma haloscope, which looks for axion excitations in a metamaterial — a solid-state plasma — should be sensitive to an axion particle of this mass, and could potentially detect one. More

Fiber reinforced composites (FRCs), which are engineering materials comprising stiff fibers embedded in a soft matrix, typically have a constant fiber radius that limits their performance. Now, researchers from the Gwangju Institute of Science and Technology in Korea have developed a scheme for AI-assisted design of FRC structures with spatially varying optimal fiber sizes, making FRCs more lightweight without compromising their mechanical strength and stiffness, which will reduce the energy consumption of cars, aircrafts, and other vehicles.
Fiber reinforced composites (FRCs) are a class of sophisticated engineering materials composed of stiff fibers embedded in a soft matrix. When properly designed, FRCs provide outstanding structural strength and stiffness for their weight, making them an attractive option for aircrafts, spacecrafts, and other vehicles where having a lightweight structure is essential.
Despite their usefulness, however, FRCs are limited by the fact that they are designed using fibers with a constant radius and a spatially-fixed fiber density, which compromises the trade-off between weight and mechanical strength. Simply put, currently available FRCs are, in fact, heavier than necessary to meet the application standards.
To tackle this issue, an international research team led by Professor Jaewook Lee of the Gwangju Institute of Science and Technology in Korea recently developed a new approach for the inverse design of FCRs with spatially-varying fiber size and orientation, also known as “functionally graded composites.” The proposed method is based on a “multiscale topology optimization,” which allows one to automatically find the best functionally graded composite structure given a set of design parameters and constraints.
“Topology optimization is an AI-based design technique that relies on computer simulation to generate an optimal structural shape instead of on the designer’s intuition and experience,” explains Prof. Lee, “On the other hand, a multiscale approach is a numerical method that combines the results of analyses conducted at different scales to derive structural characteristics.” Unlike similar existing approaches that are limited to two-dimensional functionally graded composites, the proposed methodology can simultaneously determine the optimal three-dimensional composite structure alongside its microscale fiber densities and fiber orientations.
The team demonstrated the potential of their method through several computer-assisted experiments where various functionally graded composite designs with constant or varying fiber sizes were compared. The experiments included designs for a bell crank, a displacement inverter mechanism, and a simple support beam. As expected, the results showed improved performances in the designs with locally tailored fiber sizes. This paper was made available online on October 9, 2021, and published in Volume 279 of Composite Structures on January 1, 2022.
Many applications for vehicles, aircrafts, and robotics benefit from lightweight structures, and the proposed approach can now assist engineers to this end. However, the benefits can extend well beyond the target applications themselves. As Prof. Lee explains: “Our methodology could help develop more energy-efficient vehicles and machinery by weight reduction, which would reduce their energy consumption and, in turn, contribute towards achieving carbon neutrality.”
Story Source:
Materials provided by GIST (Gwangju Institute of Science and Technology). Note: Content may be edited for style and length. More

Ultrasound scans — best known for monitoring pregnancies or imaging organs — can also be used to stimulate cells and direct cell function. A team of Penn State researchers has developed an easier, more effective way to harness the technology for biomedical applications.
The team created a transparent, biocompatible ultrasound transducer chip that resembles a microscope glass slide and can be inserted into any optical microscope for easy viewing. Cells can be cultured and stimulated directly on top of the transducer chip and the cells’ resulting changes can be imaged with optical microscopy techniques.
Published in the Royal Society of Chemistry’s journal Lab on a Chip, the paper was selected as the cover article for the December 2021 issue. Future applications of the technology could impact stem cell, cancer and neuroscience research.
“In the conventional ultrasound stimulation experiments, a cell culture dish is placed in a water bath, and a bulky ultrasound transducer directs the ultrasound waves to the cells through the water medium,” said Sri-Rajasekhar “Raj” Kothapalli, principal investigator and assistant professor of biomedical engineering at Penn State. “This was a complex setup that didn’t provide reproducible results: The results that one group saw another did not, even while using the same parameters, because there are several things that could affect the cells’ survival and stimulation while they are in water, as well as how we visualize them.”
Kothapalli and his collaborators miniaturized the ultrasound stimulation setup by creating a transparent transducer platform made of a piezoelectric lithium niobate material. Piezoelectric materials generate mechanical energy when electric voltage is applied. The chip’s biocompatible surface allows the cells to be cultured directly on the transducer and used for repeated stimulation experiments over several weeks.
When connected to a power supply, the transducer emits ultrasound waves, which pulse the cells and trigger ion influx and outflux. More

A ground-up approach to qubit design leads to a new framework for creating versatile, highly tailored quantum devices.
Advances in quantum science have the potential to revolutionize the way we live. Quantum computers hold promise for solving problems that are intractable today, and we may one day use quantum networks as hackerproof information highways.
The realization of such forward-looking technologies hinges in large part on the qubit — the fundamental component of quantum systems. A major challenge of qubit research is designing them to be customizable, tailored to work with all kinds of sensing, communication and computational devices.
Scientists have taken a major step in the development of tailored qubits. In a paper published in the Journal of the American Chemical Society, the team, which includes researchers at MIT, the University of Chicago and Columbia University, demonstrates how a particular molecular family of qubits can be finely tuned over a broad spectrum, like turning a sensitive dial on a wideband radio.
The team also outlines the underlying design features that enable exquisite control over these quantum bits.
“This is a new platform for qubit design. We can use our predictable, controllable, tunable design strategy to create a new quantum system,” said Danna Freedman, MIT professor of chemistry and a co-author of the study. ?”We’ve demonstrated the broad range of tunability over which these design principles work.”
The work is partially supported by Q-NEXT, a U.S. Department of Energy (DOE) National Quantum Information Science Research Center led by Argonne National Laboratory. More

Scintillators are materials that emit light when bombarded with high-energy particles or X-rays. In medical or dental X-ray systems, they convert incoming X-ray radiation into visible light that can then be captured using film or photosensors. They’re also used for night-vision systems and for research, such as in particle detectors or electron microscopes.
Researchers at MIT have now shown how one could improve the efficiency of scintillators by at least tenfold, and perhaps even a hundredfold, by changing the material’s surface to create certain nanoscale configurations, such as arrays of wave-like ridges. While past attempts to develop more efficient scintillators have focused on finding new materials, the new approach could in principle work with any of the existing materials.
Though it will require more time and effort to integrate their scintillators into existing X-ray machines, the team believes that this method might lead to improvements in medical diagnostic X-rays or CT scans, to reduce dose exposure and improve image quality. In other applications, such as X-ray inspection of manufactured parts for quality control, the new scintillators could enable inspections with higher accuracy or at faster speeds.
The findings are described in the journal Science, in a paper by MIT doctoral students Charles Roques-Carmes and Nicholas Rivera; MIT professors Marin Soljacic, Steven Johnson, and John Joannopoulos; and 10 others.
While scintillators have been in use for some 70 years, much of the research in the field has focused on developing new materials that produce brighter or faster light emissions. The new approach instead applies advances in nanotechnology to existing materials. By creating patterns in scintillator materials at a length scale comparable to the wavelengths of the light being emitted, the team found that it was possible to dramatically change the material’s optical properties.
To make what they coined “nanophotonic scintillators,” Roques-Carmes says, “you can directly make patterns inside the scintillators, or you can glue on another material that would have holes on the nanoscale. The specifics depend on the exact structure and material.” For this research, the team took a scintillator and made holes spaced apart by roughly one optical wavelength, or about 500 nanometers (billionths of a meter). More

Researchers from the University of Oxford’s Big Data Institute have taken a major step towards mapping the entirety of genetic relationships among humans: a single genealogy that traces the ancestry of all of us. The study has been published today in Science.
The past two decades have seen extraordinary advancements in human genetic research, generating genomic data for hundreds of thousands of individuals, including from thousands of prehistoric people. This raises the exciting possibility of tracing the origins of human genetic diversity to produce a complete map of how individuals across the world are related to each other.
Until now, the main challenges to this vision were working out a way to combine genome sequences from many different databases and developing algorithms to handle data of this size. However, a new method published today by researchers from the University of Oxford’s Big Data Institute can easily combine data from multiple sources and scale to accommodate millions of genome sequences.
Dr Yan Wong, an evolutionary geneticist at the Big Data Institute, and one of the principal authors, explained: “We have basically built a huge family tree, a genealogy for all of humanity that models as exactly as we can the history that generated all the genetic variation we find in humans today. This genealogy allows us to see how every person’s genetic sequence relates to every other, along all the points of the genome.”
Since individual genomic regions are only inherited from one parent, either the mother or the father, the ancestry of each point on the genome can be thought of as a tree. The set of trees, known as a “tree sequence” or “ancestral recombination graph,” links genetic regions back through time to ancestors where the genetic variation first appeared.
Lead author Dr Anthony Wilder Wohns, who undertook the research as part of his PhD at the Big Data Institute and is now a postdoctoral researcher at the Broad Institute of MIT and Harvard, said: “Essentially, we are reconstructing the genomes of our ancestors and using them to form a vast network of relationships. We can then estimate when and where these ancestors lived. The power of our approach is that it makes very few assumptions about the underlying data and can also include both modern and ancient DNA samples.”
The study integrated data on modern and ancient human genomes from eight different databases and included a total of 3,609 individual genome sequences from 215 populations. The ancient genomes included samples found across the world with ages ranging from 1,000s to over 100,000 years. The algorithms predicted where common ancestors must be present in the evolutionary trees to explain the patterns of genetic variation. The resulting network contained almost 27 million ancestors.
After adding location data on these sample genomes, the authors used the network to estimate where the predicted common ancestors had lived. The results successfully recaptured key events in human evolutionary history, including the migration out of Africa.
Although the genealogical map is already an extremely rich resource, the research team plans to make it even more comprehensive by continuing to incorporate genetic data as it becomes available. Because tree sequences store data in a highly efficient way, the dataset could easily accommodate millions of additional genomes.
Dr Wong said: “This study is laying the groundwork for the next generation of DNA sequencing. As the quality of genome sequences from modern and ancient DNA samples improves, the trees will become even more accurate and we will eventually be able to generate a single, unified map that explains the descent of all the human genetic variation we see today.”
Dr Wohns added: “While humans are the focus of this study, the method is valid for most living things; from orangutans to bacteria. It could be particularly beneficial in medical genetics, in separating out true associations between genetic regions and diseases from spurious connections arising from our shared ancestral history.”
Story Source:
Materials provided by University of Oxford. Note: Content may be edited for style and length. More

The field of machine learning on quantum computers got a boost from new research removing a potential roadblock to the practical implementation of quantum neural networks. While theorists had previously believed an exponentially large training set would be required to train a quantum neural network, the quantum No-Free-Lunch theorem developed by Los Alamos National Laboratory shows that quantum entanglement eliminates this exponential overhead.
“Our work proves that both big data and big entanglement are valuable in quantum machine learning. Even better, entanglement leads to scalability, which solves the roadblock of exponentially increasing the size of the data in order to learn it,” said Andrew Sornborger, a computer scientist at Los Alamos and a coauthor of the paper published Feb. 18 in Physical Review Letters. “The theorem gives us hope that quantum neural networks are on track towards the goal of quantum speed-up, where eventually they will outperform their counterparts on classical computers.”
The classical No-Free-Lunch theorem states that any machine-learning algorithm is as good as, but no better than, any other when their performance is averaged over all possible functions connecting the data to their labels. A direct consequence of this theorem that showcases the power of data in classical machine learning is that the more data one has, the better the average performance. Thus, data is the currency in machine learning that ultimately limits performance.
The new Los Alamos No-Free-Lunch theorem shows that in the quantum regime entanglement is also a currency, and one that can be exchanged for data to reduce data requirements.
Using a Rigetti quantum computer, the team entangled the quantum data set with a reference system to verify the new theorem.
“We demonstrated on quantum hardware that we could effectively violate the standard No-Free-Lunch theorem using entanglement, while our new formulation of the theorem held up under experimental test,” said Kunal Sharma, the first author on the article.
“Our theorem suggests that entanglement should be considered a valuable resource in quantum machine learning, along with big data,” said Patrick Coles, a physicist at Los Alamos and senior author on the article. “Classical neural networks depend only on big data.”
Entanglement describes the state of a system of atomic-scale particles that cannot be fully described independently or individually. Entanglement is a key component of quantum computing.
Story Source:
Materials provided by DOE/Los Alamos National Laboratory. Note: Content may be edited for style and length. More

A collaboration between Lawrence Berkeley National Laboratory’s (Berkeley Lab’s) Applied Mathematics and Computational Research Division (AMCRD) and Physics Division has yielded a new approach to error mitigation that could help make quantum computing’s theoretical potential a reality.
The research team describes this work in a paper published in Physical Review Letters, “Mitigating Depolarizing Noise on Quantum Computers with Noise-Estimation Circuits.”
“Quantum computers have the potential to solve more complex problems way faster than classical computers,” said Bert de Jong, one of the lead authors of the study and the director of the AIDE-QC and QAT4Chem quantum computing projects. De Jong also leads the AMCRD’s Applied Computing for Scientific Discovery Group. “But the real challenge is quantum computers are relatively new. And there’s still a lot of work that has to be done to make them reliable.”
For now, one of the problems is that quantum computers are still too error-prone to be consistently useful. This is due in large part to something known as “noise” (errors).
There are different types of noise, including readout noise and gate noise. The former has to do with reading out the result of a run on a quantum computer; the more noise, the higher the chance a qubit — the quantum equivalent of a bit on a classical computer — will be measured in the wrong state. The latter relates to the actual operations performed; noise here means the probability of applying the wrong operation. And the prevalence of noise dramatically increases the more operations one tries to perform with a quantum computer, which makes it harder to tease out the right answer and severely limits quantum computers’ usability as they’re scaled up.
“So noise here just basically means: It’s stuff you don’t want, and it obscures the result you do want,” said Ben Nachman, a Berkeley Lab physicist and co-author on the study who also leads the cross-cutting Machine Learning for Fundamental Physics group. More