Computers Math

Researchers at the USC Viterbi School of Engineering are using generative adversarial networks (GANs) — technology best known for creating deepfake videos and photorealistic human faces — to improve brain-computer interfaces for people with disabilities.
In a paper published in Nature Biomedical Engineering, the team successfully taught an AI to generate synthetic brain activity data. The data, specifically neural signals called spike trains, can be fed into machine-learning algorithms to improve the usability of brain-computer interfaces (BCI).
BCI systems work by analyzing a person’s brain signals and translating that neural activity into commands, allowing the user to control digital devices like computer cursors using only their thoughts. These devices can improve quality of life for people with motor dysfunction or paralysis, even those struggling with locked-in syndrome — when a person is fully conscious but unable to move or communicate.
Various forms of BCI are already available, from caps that measure brain signals to devices implanted in brain tissues. New use cases are being identified all the time, from neurorehabilitation to treating depression. But despite all of this promise, it has proved challenging to make these systems fast and robust enough for the real world.
Specifically, to make sense of their inputs, BCIs need huge amounts of neural data and long periods of training, calibration and learning.
“Getting enough data for the algorithms that power BCIs can be difficult, expensive, or even impossible if paralyzed individuals are not able to produce sufficiently robust brain signals,” said Laurent Itti, a computer science professor and study co-author. More

Companies in the digital technology industry are significantly underreporting the greenhouse gas emissions arising along the value chain of their products. Across a sample of 56 major tech companies surveyed in a study by the Technical University of Munich (TUM), more than half of these emissions were excluded from self-reporting in 2019. At approximately 390 megatons carbon dioxide equivalents, the omitted emissions are in the same ballpark as the carbon footprint of Australia. The research team has developed a method for spotting sources of error and calculating the omitted disclosures.
For policy makers and the private sector to set targets for reduced greenhouse gas emissions, it is important to know how much CO2 companies are actually emitting. However, there are no binding requirements for comprehensive accounting and full disclosure of these emissions. The Greenhouse Gas (GHG) Protocol is seen as a voluntary standard. It distinguishes three categories of emissions: Scope 1 refers to direct emissions from a company’s own activities, scope 2 refers to emissions from the production of purchased energy, and scope 3 to emissions from activities along the value chain, in other words all emissions from raw material extraction to the use of the end product. Scope 3 emissions often represent the majority of a company’s carbon footprint. Past studies have also shown that these emissions account for most reporting gaps. Until now, however, it was not possible to quantify these gaps or determine their causes.
Lena Klaaßen and Dr. Christian Stoll at the TUM School of Management of the Technical University of Munich (TUM) have developed a method for identifying reporting gaps for scope 3 emissions and used it in a case study to determine the carbon footprints of pre-selected digital technology companies. Their paper has now been published in the journal Nature Communications.
Companies publish inconsistent figures
Klaaßen and Stoll determined that many companies submit different greenhouse gas emission figures depending on where they are reporting them. They focused mainly on the companies’ own reports as compared with voluntary disclosures to the non-profit organization CDP. The annual survey of companies conducted by CDP is regarded as the most important collection of data based on the structure of the GHG Protocol. Most companies disclose lower emissions in their own reports than in the CDP survey. This could be partly due to the fact that the CDP report is intended mainly for investors, while corporate reports are addressed to the general public.
In addition, CDP leaves it up to the reporting companies to choose which of the 15 GHG Protocol categories — ranging from business travel to waste disposal — are relevant to them. The studies show that this discretionary freedom results in some companies ignoring certain categories or not fully reporting the related emissions. Most companies have reporting gaps simply because they do not receive emissions data from all suppliers and do not fill the gaps with secondary data. More

Rare earth elements from ore are vital for modern life but refining them after mining is costly, harms the environment and mostly occurs abroad.
A new study describes a proof of principle for engineering a bacterium, Gluconobacter oxydans, that takes a big first step towards meeting skyrocketing rare earth element demand in a way that matches the cost and efficiency of traditional thermochemical extraction and refinement methods and is clean enough to meet U.S. environmental standards.
“We’re trying to come up with an environmentally friendly, low-temperature, low-pressure method for getting rare earth elements out of a rock,” said Buz Barstow, the paper’s senior author and an assistant professor of biological and environmental engineering at Cornell University.
The elements — of which there are 15 in the periodic table — are necessary for everything from computers, cell phones, screens, microphones, wind turbines, electric vehicles and conductors to radars, sonars, LED lights and rechargeable batteries.
While the U.S. once refined its own rare earth elements, that production stopped more than five decades ago. Now, refinement of these elements takes place almost entirely in other countries, particularly China.
“The majority of rare earth element production and extraction is in the hands of foreign nations,” said co-author Esteban Gazel, associate professor of earth and atmospheric sciences at Cornell. “So for the security of our country and way of life, we need to get back on track to controlling that resource.”
To meet U.S. annual needs for rare earth elements, roughly 71.5 million tonnes (~78.8 million tons) of raw ore would be required to extract 10,000 kilograms (~22,000 pounds) of elements. More

An international team of researchers from the University of Minnesota Twin Cities and Kiel University in Germany have discovered a path that could lead to shape-shifting ceramic materials. This discovery could improve everything from medical devices to electronics.
The research is published open access in Nature, the world’s leading multidisciplinary science journal.
Anyone who has ever dropped a coffee cup and watched it break into several pieces, knows that ceramics are brittle. Subject to the slightest deformation, they shatter. However, ceramics are used for more than just dishes and bathroom tiles, they are used in electronics because, depending on their composition, they may be semiconducting, superconducting, ferroelectric, or insulating. Ceramics are also non corrosive and used in making a wide variety of products, including spark plugs, fiber optics, medical devices, space shuttle tiles, chemical sensors, and skis.
On the other end of the materials spectrum are shape memory alloys. They are some of the most deformable or reshapable materials known. Shape memory alloys rely on this tremendous deformability when functioning as medical stents, the backbone of a vibrant medical device industry both in the Twin Cities area and in Germany.
The origin of this shape-shifting behavior is a solid-to-solid phase transformation. Different from the process of crystallization-melting-recrystallization, crystalline solid-solid transitions take place solely in the solid state. By changing temperature (or pressure), a crystalline solid can be transformed into another crystalline solid without entering a liquid phase.
In this new research, the route to producing a reversible shape memory ceramic was anything but straightforward. The researchers first tried a recipe that has worked for the discovery of new metallic shape memory materials. That involves a delicate tuning of the distances between atoms by compositional changes, so that the two phases fit together well. They implemented this recipe, but, instead of improving the deformability of the ceramic, they observed that some specimens exploded when they passed through the phase transformation. Others gradually fell apart into a pile of powder, a phenomenon they termed “weeping.”
With yet another composition, they observed a reversible transformation, easily transforming back and forth between the phases, much like a shape memory material. The mathematical conditions under which reversible transformation occurs can be applied widely and provide a way forward toward the paradoxical shape-memory ceramic. More

Chemotherapy successfully treats many forms of cancer, but the side effects can wreak havoc on the rest of the body. Delivering drugs directly to cancer cells could help reduce these unpleasant symptoms. Now, in a proof-of-concept study, researchers reporting in ACS Nano made fish-shaped microrobots that are guided with magnets to cancer cells, where a pH change triggers them to open their mouths and release their chemotherapy cargo.
Scientists have previously made microscale (smaller than 100 µm) robots that can manipulate tiny objects, but most can’t change their shapes to perform complex tasks, such as releasing drugs. Some groups have made 4D-printed objects (3D-printed devices that change shape in response to certain stimuli), but they typically perform only simple actions, and their motion can’t be controlled remotely. In a step toward biomedical applications for these devices, Jiawen Li, Li Zhang, Dong Wu and colleagues wanted to develop shape-morphing microrobots that could be guided by magnets to specific sites to deliver treatments. Because tumors exist in acidic microenvironments, the team decided to make the microrobots change shape in response to lowered pH.
So the researchers 4D printed microrobots in the shape of a crab, butterfly or fish using a pH-responsive hydrogel. By adjusting the printing density at certain areas of the shape, such as the edges of the crab’s claws or the butterfly’s wings, the team encoded pH-responsive shape morphing. Then, they made the microrobots magnetic by placing them in a suspension of iron oxide nanoparticles.
The researchers demonstrated various capabilities of the microrobots in several tests. For example, a fish-shaped microrobot had an adjustable “mouth” that opened and closed. The team showed that they could steer the fish through simulated blood vessels to reach cancer cells at a specific region of a petri dish. When they lowered the pH of the surrounding solution, the fish opened its mouth to release a chemotherapy drug, which killed nearby cells. Although this study is a promising proof of concept, the microrobots need to be made even smaller to navigate actual blood vessels, and a suitable imaging method needs to be identified to track their movements in the body, the researchers say.
The authors acknowledge funding from the National Natural Science Foundation of China, the National Key R&D Program of China, Major Scientific and Technological Projects in Anhui Province, the Fundamental Research Funds for the Central Universities, the Youth Innovation Promotion Association of the Chinese Academy of Sciences, the Hong Kong Research Grants Council, CAS-Croucher Funding Scheme for Joint Laboratories, the Hong Kong Special Administrative Region of the People’s Republic of China Innovation and Technology Commission and the Multi-scale Medical Robotics Center.
Story Source:
Materials provided by American Chemical Society. Note: Content may be edited for style and length. More

Northwestern University researchers have invented a new high-resolution camera that can see the unseen — including around corners and through scattering media, such as skin, fog or potentially even the human skull.
Called synthetic wavelength holography, the new method works by indirectly scattering coherent light onto hidden objects, which then scatters again and travels back to a camera. From there, an algorithm reconstructs the scattered light signal to reveal the hidden objects. Due to its high temporal resolution, the method also has potential to image fast-moving objects, such as the beating heart through the chest or speeding cars around a street corner.
The study will be published on Nov. 17 in the journal Nature Communications.
The relatively new research field of imaging objects behind occlusions or scattering media is called non-line-of-sight (NLoS) imaging. Compared to related NLoS imaging technologies, the Northwestern method can rapidly capture full-field images of large areas with submillimeter precision. With this level of resolution, the computational camera could potentially image through the skin to see even the tiniest capillaries at work.
While the method has obvious potential for noninvasive medical imaging, early-warning navigation systems for automobiles and industrial inspection in tightly confined spaces, the researchers believe potential applications are endless.
“Our technology will usher in a new wave of imaging capabilities,” said Northwestern’s Florian Willomitzer, first author of the study. “Our current sensor prototypes use visible or infrared light, but the principle is universal and could be extended to other wavelengths. For example, the same method could be applied to radio waves for space exploration or underwater acoustic imaging. It can be applied to many areas, and we have only scratched the surface.”
Willomitzer is a research assistant professor of electrical and computer engineering at Northwestern’s McCormick School of Engineering. Northwestern co-authors include Oliver Cossairt, associate professor of computer science and electrical and computer engineering, and former Ph.D. student Fengqiang Li. The Northwestern researchers collaborated closely with Prasanna Rangarajan, Muralidhar Balaji and Marc Christensen, all researchers at Southern Methodist University. More

Physicists from Exeter and Zaragoza have created a theory describing how non-reciprocity can be induced at the quantum level, paving the way for non-reciprocal transport in the next generation of nanotechnology.
A pair of theoretical physicists, from the University of Exeter (United Kingdom) and the University of Zaragoza (Spain), have developed a quantum theory explaining how to engineer non-reciprocal flows of quantum light and matter. The research may be important for the creation of quantum technologies which require the directional transfer of energy and information at small scales.
Reciprocity, going the same way backward as forward, is a ubiquitous concept in physics. A famous example may be found in Newton’s Law: for every action there is an equal and opposite reaction. The breakdown of such a powerful notion as reciprocity in any area of physics, from mechanics to optics to electromagnetism, is typically associated with surprises which can be exploited for technological application. For example, a nonreciprocal electric diode allows current to pass in forwards but not backwards and forms a building block of the microelectronics industry.
In their latest research, Downing and Zueco provide a quantum theory of non-reciprocal transport around a triangular cluster of strongly interacting quantum objects. Inspired by the physics of quantum rings, they show that by engineering an artificial magnetic field one may tune the direction of the energy flow around the cluster. The theory accounts for strong particle interactions, such that directionality appears at a swathe of energies, and considers the pernicious effect of dissipation for the formation of non-reciprocal quantum currents.
The research may be useful in the development of quantum devices requiring efficient, directional transportation, as well for further studies of strongly interacting quantum phases, synthetic magnetic fields and quantum simulators.
Charles Downing from the University of Exeter explains: “Our calculations provide insight into how one may instigate directional transport in closed nanoscopic lattices of atoms and photons with strong interactions, which may lead to the development of novel devices of a highly directional character.”
“Non-reciprocal population dynamics in a quantum trimer” is published in Proceedings of the Royal Society A, a historic journal which has been publishing scientific research since 1905.
Story Source:
Materials provided by University of Exeter. Note: Content may be edited for style and length. More

UT Southwestern and University of Washington researchers led an international team that used artificial intelligence (AI) and evolutionary analysis to produce 3D models of eukaryotic protein interactions. The study, published in Science, identified more than 100 probable protein complexes for the first time and provided structural models for more than 700 previously uncharacterized ones. Insights into the ways pairs or groups of proteins fit together to carry out cellular processes could lead to a wealth of new drug targets.
“Our results represent a significant advance in the new era in structural biology in which computation plays a fundamental role,” said Qian Cong, Ph.D., Assistant Professor in the Eugene McDermott Center for Human Growth and Development with a secondary appointment in Biophysics.
Dr. Cong led the study with David Baker, Ph.D., Professor of Biochemistry and Dr. Cong’s postdoctoral mentor at the University of Washington prior to her recruitment to UT Southwestern. The study has four co-lead authors, including UT Southwestern Computational Biologist Jimin Pei, Ph.D.
Proteins often operate in pairs or groups known as complexes to accomplish every task needed to keep an organism alive, Dr. Cong explained. While some of these interactions are well studied, many remain a mystery. Constructing comprehensive interactomes — or descriptions of the complete set of molecular interactions in a cell — would shed light on many fundamental aspects of biology and give researchers a new starting point on developing drugs that encourage or discourage these interactions. Dr. Cong works in the emerging field of interactomics, which combines bioinformatics and biology.
Until recently, a major barrier for constructing an interactome was uncertainty over the structures of many proteins, a problem scientists have been trying to solve for half a century. In 2020 and 2021, a company called DeepMind and Dr. Baker’s lab independently released two AI technologies called AlphaFold (AF) and RoseTTAFold (RF) that use different strategies to predict protein structures based on the sequences of the genes that produce them.
In the current study, Dr. Cong, Dr. Baker, and their colleagues expanded on those AI structure-prediction tools by modeling many yeast protein complexes. Yeast is a common model organism for fundamental biological studies. To find proteins that were likely to interact, the scientists first searched the genomes of related fungi for genes that acquired mutations in a linked fashion. They then used the two AI technologies to determine whether these proteins could be fit together in 3D structures. More