Subterms
Computer scientists developed a deep learning method to create realistic objects for virtual environments that can be used to train robots. The researchers used TACC’s Maverick2 supercomputer to train the generative adversarial network. The network is the first that can produce colored point clouds with fine details at multiple resolutions. More
Following a series of studies on termite mound physiology and morphogenesis over the past decade, researchers have now developed a mathematical model to help explain how termites construct their intricate mounds. More
Scientists have successfully used satellite cameras coupled with deep learning to count animals in complex geographical landscapes, taking conservationists an important step forward in monitoring populations of endangered species. More
Researchers present an optical flow-based learning process that allows robots to estimate distances through the visual appearance (shape, color, texture) of the objects in view. This artificial intelligence (AI)-based learning strategy increases the navigation skills of small flying drones and entails a new hypothesis on insect intelligence. More
Quantum nanowires — which have length but no width or height-provide a unique environment for the formation and detection of a quasiparticle known as a Majorana zero mode.
A new UNSW-led study overcomes previous difficulty detecting the Majorana zero mode, and produces a significant improvement in device reproducibility.
Potential applications for Majorana zero modes include fault-resistant topological quantum computers, and topological superconductivity.
MAJORANA FERMIONS IN 1D WIRES
A Majorana fermion is a composite particle that is its own antiparticle.
advertisement
Antimatter explainer: Every fundamental particle has a corresponding antimatter particle, with the same mass but opposite electrical charge. For example, the antiparticle of an electron (charge -1) is a positron (charge +1)
Such unusual particle’s interest academically and commercially comes from their potential use in a topological quantum computer, predicted to be immune to the decoherence that randomises the precious quantum information.
Majorana zero modes can be created in quantum wires made from special materials in which there is a strong coupling between their electrical and magnetic properties.
In particular, Majorana zero modes can be created in one-dimensional semiconductors (such as semiconductor nanowires) when coupled with a superconductor.
In a one-dimensional nanowire, whose dimensions perpendicular to length are small enough not to allow any movement of subatomic particles, quantum effects predominate.
advertisement
NEW METHOD FOR DETECTING NECESSARY SPIN-ORBIT GAP
Majorana fermions, which are their own antiparticle, have been theorised since 1937, but have only been experimentally observed in the last decade. The Majorana fermion’s ‘immunity’ to decoherence provides potential use for fault-tolerant quantum computing.
One-dimensional semiconductor systems with strong spin-orbit interaction are attracting great attention due to potential applications in topological quantum computing.
The magnetic ‘spin’ of an electron is like a little bar magnet, whose orientation can be set with an applied magnetic field.
In materials with a ‘spin-orbit interaction’ the spin of an electron is determined by the direction of motion, even at zero magnetic field. This allows for all electrical manipulation of magnetic quantum properties.
Applying a magnetic field to such a system can open an energy gap such that forward -moving electrons all have the same spin polarisation, and backward-moving electrons have the opposite polarisation. This ‘spin-gap’ is a pre-requisite for the formation of Majorana zero modes.
Despite intense experimental work, it has proven extremely difficult to unambiguously detect this spin-gap in semiconductor nanowires, since the spin-gap’s characteristic signature (a dip in its conductance plateau when a magnetic field is applied) is very hard to distinguish from unavoidable the background disorder in nanowires.
The new study finds a new, unambiguous signature for the spin-orbit gap that is impervious to the disorder effects plaguing previous studies.
“This signature will become the de-facto standard for detecting spin-gaps in the future,” says lead author Dr Karina Hudson.
REPRODUCIBILITY
The use of Majorana zero modes in a scalable quantum computer faces an additional challenge due to the random disorder and imperfections in the self-assembled nanowires that host the MZM.
It has previously been almost impossible to fabricate reproducible devices, with only about 10% of devices functioning within desired parameters.
The latest UNSW results show a significant improvement, with reproducible results across six devices based on three different starting wafers.
“This work opens a new route to making completely reproducible devices,” says corresponding author Prof Alex Hamilton UNSW). More
More people could be protected from life-threatening rabies thanks to an agile approach to dog vaccination using smart phone technology to spot areas of low vaccination coverage in real time. The work could help save the lives of children worldwide. More
Mount Sinai researchers have published one of the first studies using a machine learning technique called “federated learning” to examine electronic health records to better predict how COVID-19 patients will progress. The study was published in the Journal of Medical Internet Research — Medical Informatics on January 18.
The researchers said the emerging technique holds promise to create more robust machine learning models that extend beyond a single health system without compromising patient privacy. These models, in turn, can help triage patients and improve the quality of their care.
Federated learning is a technique that trains an algorithm across multiple devices or servers holding local data samples but avoids clinical data aggregation, which is undesirable for reasons including patient privacy issues. Mount Sinai researchers implemented and assessed federated learning models using data from electronic health records at five separate hospitals within the Health System to predict mortality in COVID-19 patients. They compared the performance of a federated model against ones built using data from each hospital separately, referred to as local models. After training their models on a federated network and testing the data of local models at each hospital, the researchers found the federated models demonstrated enhanced predictive power and outperformed local models at most of the hospitals.
“Machine learning models in health care often require diverse and large-scale data to be robust and translatable outside the patient population they were trained on,” said the study’s corresponding author, Benjamin Glicksberg, PhD, Assistant Professor of Genetics and Genomic Sciences at the Icahn School of Medicine at Mount Sinai, and member of the Hasso Plattner Institute for Digital Health at Mount Sinai and the Mount Sinai Clinical Intelligence Center. “Federated learning is gaining traction within the biomedical space as a way for models to learn from many sources without exposing any sensitive patient data. In our work, we demonstrate that this strategy can be particularly useful in situations like COVID-19.”
Machine learning models built within a hospital are not always effective for other patient populations, partially due to models being trained on data from a single group of patients which is not representative of the entire population.
“Machine learning in health care continues to suffer a reproducibility crisis,” said the study’s first author, Akhil Vaid, MD, postdoctoral fellow in the Department of Genetics and Genomic Sciences at the Icahn School of Medicine at Mount Sinai, and member of the Hasso Plattner Institute for Digital Health at Mount Sinai and the Mount Sinai Clinical Intelligence Center. “We hope that this work showcases benefits and limitations of using federated learning with electronic health records for a disease that has a relative dearth of data in an individual hospital. Models built using this federated approach outperform those built separately from limited sample sizes of isolated hospitals. It will be exciting to see the results of larger initiatives of this kind.”
Story Source:
Materials provided by The Mount Sinai Hospital / Mount Sinai School of Medicine. Note: Content may be edited for style and length. More
A student infected with COVID-19 returning home from university for Christmas would, on average, have infected just less than one other household member with the virus, according to a new model devised by mathematicians at Cardiff University and published in Health Systems.
Professor Paul Harper and colleagues defined an equation to predict the number of secondary household infections using variables for prevalence of the virus, the probability of secondary transmission, the number of household occupants and the total number of students returning home.
The model predicts that each infected student allowed to return home would produce, on average, 0.94 secondary infections.
“With the potential movement of over 1 million UK students for the Christmas vacation, even a modest 1% infection level (meaning 10 in 1,000 students are infected, perhaps many of them without symptoms at the time of travel) would equate to 9,400 new secondary household cases across the country,” says Professor Harper.
As the study does not consider transmission to the students’ wider home communities or include the journey home — which may give rise to a larger number of cases, particularly if public transport is taken — the numbers are a lower bound on the likely impact of transmissions and new cases.
However, although the indicative levels of secondary infections are potentially very large, multiple strategies can be adopted to help reduce the number of students taking Covid-19 home, the authors say. These include strongly advising students not to mix in the days leading up to departure, implementing staggered departure times and facilitating mass testing of students before they head home.
The authors have provided computer code and an online app to allow anyone to rerun and adapt the simulations. “The code and app are quick to run with a focus on accessibility so that a user can rapidly change the input probabilities to suit their data, thereby generating their own results based on localised parameters,” the authors say.
Their results have been presented to the Welsh Government and have informed policy in relation to the two-week firebreak in Wales in October/November and for the forthcoming vacation. The data has also been communicated across the governments of England, Scotland and Northern Ireland.
Story Source:
Materials provided by Taylor & Francis Group. Note: Content may be edited for style and length. More
