Computers Math

Today, researchers are developing ways to accelerate discovery by combining automated experiments, artificial intelligence and high-performance computing. A novel tool developed at Oak Ridge National Laboratory that leverages those technologies has demonstrated that AI can influence materials synthesis and conduct associated experiments without human supervision.
This autonomous materials synthesis tool uses pulsed laser deposition, or PLD, to deposit a thin layer of substance onto a base material. It then employs AI to analyze how the quality of the newly created material relates to the synthesis conditions, such as temperature, pressure and energy emitted during the PLD process. The AI suggests a revised set of conditions that may yield improved quality and then controls the PLD equipment to conduct the next experiment.
“We built computer control of all processes into the system and incorporated some hardware innovations to enable AI to drive experimentation,” said the study’s leader, Sumner Harris of the Center for Nanophase Materials Sciences at ORNL. “The automation allows us to perform our work 10 times faster, and the AI can understand huge parameter spaces with far fewer samples.” More

The need to transition away from fossil fuels to more sustainable energy production is critical. That’s why a team of Concordia researchers is looking at a potential power source that not only produces no carbon emissions but removes carbon as it works: algae.
Researchers at the Optical-Bio Microsystems Lab recently published a new paper on this topic in the journal Energies. In it, they describe their method of extracting energy from the photosynthesis process of algae suspended in a specialized solution and housed in small power cells. Configured correctly, these cells can generate enough energy to power low- and ultra-low power devices such as Internet of Things (IoT) sensors.
“The idea of the micro photosynthetic power cell is to extract electrons produced through the process of photosynthesis,” says Kirankumar Kuruvinashetti, PhD 20, now a Mitacs postdoctoral associate at the University of Calgary.
“Photosynthesis produces oxygen and electrons. Our model traps the electrons, which allows us to generate electricity. So more than being a zero-emission technology, it’s a negative carbon emission technology: it absorbs carbon dioxide from the atmosphere and gives you a current. Its only byproduct is water.”
Power generated day and night
The micro photosynthetic power cell consists of an anode and a cathode chamber separated by a honeycomb-shaped proton exchange membrane. The researchers fabricated microelectrodes on both sides of the membrane to collect the charges released by the algae during photosynthesis. Each chamber measures only two centimetres by two centimetres by four millimetres.
The algae are suspended in a two-millilitre solution in the anode chamber while the cathode is filled with potassium ferricyanide, a type of electron acceptor. Once the algae undergo photosynthesis and begin to release electrons, the electrons will be collected through the membrane’s electrodes and conducted, creating a current.

The protons, meanwhile, will pass through the membrane into the cathode and cause oxidation, resulting in a potassium ferrocyanide reduction.
The process also works without direct sunlight, though at a lower intensity, explains PhD candidate and paper co-author Dhilippan Panneerselvam.
“Just like humans, algae are constantly breathing — but they intake carbon dioxide and release oxygen. Due to their photosynthesis machinery, they also release electrons during respiration. The electricity generation is not stopped. The electrons are continuously harvested.”
Muthukumaran Packirisamy, professor in the Department of Mechanical, Industrial and Aerospace Engineering and the paper’s corresponding author, admits the system is not yet able to compete in power generation with others like photovoltaic cells. The maximum possible terminal voltage of a single micro photosynthetic power cell is only 1.0V.
But he believes that, with enough research and development, including artificial intelligence-assisted integration technologies, this technology has the potential to be a viable, affordable and clean power source in the future.
It also offers significant manufacturing advantages over other systems, he says.
“Our system does not use any of the hazardous gases or microfibres needed for the silicon fabrication technology that photovoltaic cells rely on. Furthermore, disposing of silicon computer chips is not easy. We use biocompatible polymers, so the whole system is easily decomposable and very cheap to manufacture.” More

The agility with which humans and animals move is an evolutionary marvel that no robot has yet been able to closely emulate. To help probe the mystery of how brains control movement, Harvard neuroscientists have created a virtual rat with an artificial brain that can move around just like a real rodent.
Bence Ölveczky, professor in the Department of Organismic and Evolutionary Biology, led a group of researchers who collaborated with scientists at Google’s DeepMind AI lab to build a biomechanically realistic digital model of a rat. Using high-resolution data recorded from real rats, they trained an artificial neural network — the virtual rat’s “brain” — to control the virtual body in a physics simulator called MuJoco, where gravity and other forces are present.
Publishing in Nature, the researchers found that activations in the virtual control network accurately predicted neural activity measured from the brains of real rats producing the same behaviors, said Ölveczky, who is an expert at training (real) rats to learn complex behaviors in order to study their neural circuitry. The feat represents a new approach to studying how the brain controls movement, Ölveczky said, by leveraging advances in deep reinforcement learning and AI, as well as 3D movement-tracking in freely behaving animals.
The collaboration was “fantastic,” Ölveczky said. “DeepMind had developed a pipeline to train biomechanical agents to move around complex environments. We simply didn’t have the resources to run simulations like those, to train these networks.”
Working with the Harvard researchers was, likewise, “a really exciting opportunity for us,” said co-author and Google DeepMind Senior Director of Research Matthew Botvinick. “We’ve learned a huge amount from the challenge of building embodied agents: AI systems that not only have to think intelligently, but also have to translate that thinking into physical action in a complex environment. It seemed plausible that taking this same approach in a neuroscience context might be useful for providing insights in both behavior and brain function.”
Graduate student Diego Aldarondo worked closely with DeepMind researchers to train the artificial neural network to implement what are called inverse dynamics models, which scientists believe our brains use to guide movement. When we reach for a cup of coffee, for example, our brain quickly calculates the trajectory our arm should follow and translates this into motor commands. Similarly, based on data from actual rats, the network was fed a reference trajectory of the desired movement and learned to produce the forces to generate it. This allowed the virtual rat to imitate a diverse range of behaviors, even ones it hadn’t been explicitly trained on.
These simulations may launch an untapped area of virtual neuroscience in which AI-simulated animals, trained to behave like real ones, provide convenient and fully transparent models for studying neural circuits, and even how such circuits are compromised in disease. While Ölveczky’s lab is interested in fundamental questions about how the brain works, the platform could be used, as one example, to engineer better robotic control systems.
A next step might be to give the virtual animal autonomy to solve tasks akin to those encountered by real rats. “From our experiments, we have a lot of ideas about how such tasks are solved, and how the learning algorithms that underlie the acquisition of skilled behaviors are implemented,” Ölveczky continued. “We want to start using the virtual rats to test these ideas and help advance our understanding of how real brains generate complex behavior.” More

Quantum computers have the potential to solve complex problems in human health, drug discovery, and artificial intelligence millions of times faster than some of the world’s fastest supercomputers. A network of quantum computers could advance these discoveries even faster. But before that can happen, the computer industry will need a reliable way to string together billions of qubits — or quantum bits — with atomic precision.
Connecting qubits, however, has been challenging for the research community. Some methods form qubits by placing an entire silicon wafer in a rapid annealing oven at very high temperatures. With these methods, qubits randomly form from defects (also known as color centers or quantum emitters) in silicon’s crystal lattice. And without knowing exactly where qubits are located in a material, a quantum computer of connected qubits will be difficult to realize.
But now, getting qubits to connect may soon be possible. A research team led by Lawrence Berkeley National Laboratory (Berkeley Lab) says that they are the first to use a femtosecond laser to create and “annihilate” qubits on demand, and with precision, by doping silicon with hydrogen.
The advance could enable quantum computers that use programmable optical qubits or “spin-photon qubits” to connect quantum nodes across a remote network. It could also advance a quantum internet that is not only more secure but could also transmit more data than current optical-fiber information technologies.
“To make a scalable quantum architecture or network, we need qubits that can reliably form on-demand, at desired locations, so that we know where the qubit is located in a material. And that’s why our approach is critical,” said Kaushalya Jhuria, a postdoctoral scholar in Berkeley Lab’s Accelerator Technology & Applied Physics (ATAP) Division. She is the first author on a new study that describes the technique in the journal Nature Communications. “Because once we know where a specific qubit is sitting, we can determine how to connect this qubit with other components in the system and make a quantum network.”
“This could carve out a potential new pathway for industry to overcome challenges in qubit fabrication and quality control,” said principal investigator Thomas Schenkel, head of the Fusion Science & Ion Beam Technology Program in Berkeley Lab’s ATAP Division. His group will host the first cohort of students from the University of Hawaii in June as part of a DOE Fusion Energy Sciences-funded RENEW project on workforce development where students will be immersed in color center/qubit science and technology.
Forming qubits in silicon with programmable control
The new method uses a gas environment to form programmable defects called “color centers” in silicon. These color centers are candidates for special telecommunications qubits or “spin photon qubits.” The method also uses an ultrafast femtosecond laser to anneal silicon with pinpoint precision where those qubits should precisely form. A femtosecond laser delivers very short pulses of energy within a quadrillionth of a second to a focused target the size of a speck of dust.

Spin photon qubits emit photons that can carry information encoded in electron spin across long distances — ideal properties to support a secure quantum network. Qubits are the smallest components of a quantum information system that encodes data in three different states: 1, 0, or a superposition that is everything between 1 and 0.
With help from Boubacar Kanté, a faculty scientist in Berkeley Lab’s Materials Sciences Division and professor of electrical engineering and computer sciences (EECS) at UC Berkeley, the team used a near-infrared detector to characterize the resulting color centers by probing their optical (photoluminescence) signals.
What they uncovered surprised them: a quantum emitter called the Ci center. Owing to its simple structure, stability at room temperature, and promising spin properties, the Ci center is an interesting spin photon qubit candidate that emits photons in the telecom band. “We knew from the literature that Ci can be formed in silicon, but we didn’t expect to actually make this new spin photon qubit candidate with our approach,” Jhuria said.
The researchers learned that processing silicon with a low femtosecond laser intensity in the presence of hydrogen helped to create the Ci color centers. Further experiments showed that increasing the laser intensity can increase the mobility of hydrogen, which passivates undesirable color centers without damaging the silicon lattice, Schenkel explained.
A theoretical analysis performed by Liang Tan, staff scientist in Berkeley Lab’s Molecular Foundry, shows that the brightness of the Ci color center is boosted by several orders of magnitude in the presence of hydrogen, confirming their observations from laboratory experiments.
“The femtosecond laser pulses can kick out hydrogen atoms or bring them back, allowing the programmable formation of desired optical qubits in precise locations,” Jhuria said.

The team plans to use the technique to integrate optical qubits in quantum devices such as reflective cavities and waveguides, and to discover new spin photon qubit candidates with properties optimized for selected applications.
“Now that we can reliably make color centers, we want to get different qubits to talk to each other — which is an embodiment of quantum entanglement — and see which ones perform the best. This is just the beginning,” said Jhuria.
“The ability to form qubits at programmable locations in a material like silicon that is available at scale is an exciting step towards practical quantum networking and computing,” said Cameron Geddes, Director of the ATAP Division.
Theoretical analysis for the study was performed at the Department of Energy’s National Energy Research Scientific Computing Center (NERSC) at Berkeley Lab with support from the NERSC QIS@Perlmutter program.
The Molecular Foundry and NERSC are DOE Office of Science user facilities at Berkeley Lab.
This work was supported by the DOE Office of Fusion Energy Sciences. More

Today’s intelligent robots can accurately recognize many objects through vision and touch. Tactile information, obtained through sensors, along with machine learning algorithms, enables robots to identify objects previously handled.
However, sensing is often confused when presented with objects similar in size and shape, or objects unknown to the robot. Other factors restrictive to robot perception include background noise and the same type of object with different shapes and sizes.
In Applied Physics Reviews, by AIP Publishing, researchers from Tsinghua University worked to break through the difficulties of robotic recognition of various common, yet complex, items.
Humans possess many different types of touch sensing, one of which is thermal feeling. This allows us to sense the wind blowing, perceive hot and cold, and discriminate between matter types, such as wood and metal, because of the different cooling sensations produced. The researchers aimed to mimic this ability by designing a robotic tactile sensing method that incorporated thermal sensations for more robust and accurate object detection.
“We propose utilizing spatiotemporal tactile sensing during hand grasping to extend the robotic function and ability to simultaneously perceive multi-attributes of the grasped object, including thermal conductivity, thermal diffusivity, surface roughness, contact pressure, and temperature,” said author Rong Zhu.
The team created a layered sensor with material detection at the surface and pressure sensitivity at the bottom, with a porous middle layer sensitive to thermal changes. They paired this sensor with an efficient cascade classification algorithm that rules out object types in order, from easy to hard, starting with simple categories like empty cartons before moving on to orange peels or scraps of cloth.
To test the capabilities of their method, the team created an intelligent robot tactile system to sort garbage. The robot picked up a range of common trash items, including empty cartons, bread scraps, plastic bags, plastic bottles, napkins, sponges, orange peels, and expired drugs. It sorted the trash into separate containers for recyclables, food scraps, hazardous waste, and other waste. Their system achieved a classification accuracy of 98.85% in recognizing diverse garbage objects not encountered previously. This successful garbage sorting behavior could greatly reduce human labor in real-life scenarios and provide a broad applicability for smart life technologies.
Future research in this area will focus on enhancing robotic embodied intelligence and autonomous implementation.
“In addition, by combining this sensor with brain-computer interface technology, tactile information collected by the sensor could be converted into neural signals acceptable to the human brain, re-empowering tactile perception capabilities for people with hand disabilities,” said Zhu. More

A study led by Nagoya University in Japan revealed that a simple thermal reaction of gallium nitride (GaN) with metallic magnesium (Mg) results in the formation of a distinctive superlattice structure. This represents the first time researchers have identified the insertion of 2D metal layers into a bulk semiconductor. By carefully observing the materials through various cutting-edge characterization techniques, the researchers uncovered new insights into the process of semiconductor doping and elastic strain engineering. They published their findings in the journal Nature.
GaN is an important wide bandgap semiconductor material that is poised to replace traditional silicon semiconductors in applications demanding higher power density and faster operating frequencies. These distinctive characteristics of GaN make it valuable in devices such as LEDs, laser diodes, and power electronics — including critical components in electric vehicles and fast chargers. The improved performance of GaN-based devices contributes to the realization of an energy-saving society and a carbon-neutral future.
In semiconductors, there are two essential and complementary types of electrical conductivity: p-type and n-type. The p-type semiconductor features primarily free carriers carrying positive charges, known as holes, whereas the n-type semiconductor conducts electricity through free electrons.
A semiconductor acquires p-type or n-type conductivity through a process called doping, which refers to the intentional introduction of specific impurities (known as dopants) into a pure semiconductor material to greatly alter its electrical and optical properties.
In the field of GaN semiconductors, Mg is the only known element to create p-type conductivity up to now. Despite 35 years since the first success of doping Mg into GaN, the full mechanisms of Mg doping in GaN, especially the solubility limit and segregation behavior of Mg, remain unclear. This uncertainty limits their optimization for optoelectronics and electronics.
To improve the conductivity of p-type GaN, Jia Wang, the first author of the study, and his colleagues conducted an experiment in which they patterned deposited metallic Mg thin films on GaN wafers and heated them up at a high temperature — a conventional process known as annealing.
Using state-of-the-art electron microscope imaging, the scientists observed the spontaneous formation of a superlattice featuring alternating layers of GaN and Mg. This is especially unusual since GaN and Mg are two types of materials with significant differences in their physical properties.
“Although GaN is a wide-bandgap semiconductor with mixed ionic and covalent bonding, and Mg is a metal featuring metallic bonding, these two dissimilar materials have the same crystal structure, and it is a strikingly natural coincidence that the lattice difference between hexagonal GaN and hexagonal Mg is negligibly small,” Wang said. “We think that the perfect lattice match between GaN and Mg greatly reduces the energy needed to create the structure, playing a critical role in the spontaneous formation of such a superlattice.”
The researchers determined that this unique intercalation behavior, which they named interstitial intercalation, leads to compressive strain to the host material. Specifically, they found that the GaN being inserted with Mg layers sustains a high stress of more than 20 GPa, equivalent to 200,000 times atmospheric pressure, making it the highest compressive strain ever recorded in a thin-film material. This is much more than the compressive stresses commonly found in silicon films (in the range of 0.1 to 2 GPa). Electronic thin films can undergo significant changes in electronic and magnetic properties because of this strain. The researchers found that the electrical conductivity in GaN via hole transport was significantly enhanced along the strained direction.
“Using such a simple and low-cost approach, we were able to enhance the transport of holes in GaN, which conducts more current,” Wang said. “This interesting finding in interactions between a semiconductor and a metal may provide new insights into semiconductor doping and improve the performance of GaN-based devices.” More

Physicists at TU Graz have calculated how suitable molecules can be stimulated by infrared light pulses to form tiny magnetic fields. If this is also successful in experiments, the principle could be used in quantum computer circuits.
When molecules are irradiated with infrared light, they begin to vibrate due to the energy supply. For Andreas Hauser from the Institute of Experimental Physics at Graz University of Technology (TU Graz), this well-known phenomenon was the starting point for considering whether these oscillations could also be used to generate magnetic fields. This is because atomic nuclei are positively charged, and when a charged particle moves, a magnetic field is created. Using the example of metal phthalocyanines — ring-shaped, planar dye molecules — Andreas Hauser and his team have now calculated that, due to their high symmetry, these molecules actually generate tiny magnetic fields in the nanometre range when infrared pulses act on them. According to the calculations, it should be possible to measure the rather low but very precisely localised field strength using nuclear magnetic resonance spectroscopy. The researchers have published their results in the Journal of the American Chemical Society.
Circular dance of the molecules
For the calculations, the team drew on preliminary work from the early days of laser spectroscopy, some of which was decades old, and used modern electron structure theory on supercomputers at the Vienna Scientific Cluster and TU Graz to calculate how phthalocyanine molecules behave when irradiated with circularly polarised infrared light. What happened was that the circularly polarised, i.e. helically twisted, light waves excite two molecular vibrations at the same time at right angles to each other. “As every rumba dancing couple knows, the right combination of forwards-backwards and left-right creates a small, closed loop. And this circular movement of each affected atomic nucleus actually creates a magnetic field, but only very locally, with dimensions in the range of a few nanometres,” says Andreas Hauser.
Molecules as circuits in quantum computers
By selectively manipulating the infrared light, it is even possible to control the strength and direction of the magnetic field, explains Andreas Hauser. This would turn the molecules into high-precision optical switches, which could perhaps also be used to build circuits for a quantum computer.
Experiments as next step
Together with colleagues from the Institute of Solid State Physics at TU Graz and a team at the University of Graz, Andreas Hauser now wants to prove experimentally that molecular magnetic fields can be generated in a controlled manner. “For proof, but also for future applications, the phthalocyanine molecule needs to be placed on a surface. However, this changes the physical conditions, which in turn influences the light-induced excitation and the characteristics of the magnetic field,” explains Andreas Hauser. “We therefore want to find a support material that has minimal impact on the desired mechanism.” In a next step, the physicist and his colleagues want to compute the interactions between the deposited phthalocyanines, the support material and the infrared light before putting the most promising variants to the test in experiments. More

A computer program based on data from nearly a half-million tissue images and powered by artificial intelligence can accurately diagnose cases of adenocarcinoma, the most common form of lung cancer, a new study shows.
Researchers at NYU Langone Health’s Perlmutter Cancer Center and the University of Glasgow developed and tested the program. They say that because it incorporates structural features of tumors from 452 adenocarcinoma patients, who are among the more than 11,000 patients in the United States National Cancer Institute’s Cancer Genome Atlas, the program offers an unbiased, detailed, and reliable second opinion for patients and oncologists about the presence of the cancer and the likelihood and timing of its return (prognosis).
The research team also points out that the program is independent and “self-taught,” meaning that it determined on its own which structural features were statistically most significant to gauging the severity of disease and had the greatest impact on tumor recurrence.
Publishing in the journal Nature Communications online June 11, the study program, also called an algorithm, or specifically, histomorphological phenotype learning (HPL), was found to accurately distinguish between similar lung cancers, adenocarcinoma and squamous cell cancers, 99% of the time. The HPL program was also found to be 72% accurate at predicting the likelihood and timing of cancer’s return after therapy, bettering the 64% accuracy in the predictions made by pathologists who directly examined the same patients’ tumor images, researchers say.
“Our new histomorphological phenotype learning program has the potential to offer cancer specialists and their patients a quick and unbiased diagnostic tool for lung adenocarcinoma that, once further testing is complete, can also be used to help validate and even guide their treatment decisions,” said study lead investigator Nicolas Coudray, PhD, a bioinformatics programmer at NYU Grossman School of Medicine and Perlmutter Cancer Center.
“Patients, physicians, and researchers know they can rely on this predictive modeling because it is self-taught, provides explainable decisions, and is based only on the knowledge drawn specifically from each patient’s tissue, including such features as its proportion of dying cells, tumor-fighting immune cells, and how densely packed the tumor cells are, among other features,” said Coudray.
“Lung tissue samples can now be analyzed in minutes by our computer program to provide fairly accurate predictions of whether their cancer will return, predictions that are better than current standards of care for making a prognosis in lung adenocarcinoma,” said study co-senior investigator Aristotelis Tsirigos, PhD. Tsirigos is a professor in the Departments of Pathology and Medicine at NYU Grossman School of Medicine and Perlmutter Cancer Center, where he also serves as co-director of precision medicine and director of its Applied Bioinformatics Laboratories.

Tsirigos says that thanks to such tools and other advances in the lung cancer biology, pathologists will be examining tissue scans on their computer screens, and less and less on microscopes, and then using their AI program to analyze the image and produce its own image of the scan. The new image, or “landscape,” they add, will offer a detailed breakdown of the tissue’s content. It might note, for example, that there is 5% necrosis and 10% tumor infiltration and what that means in terms of survival. That reading may statistically equate to an 80% chance of remaining cancer-free for two years or more, based on information from all the patient data in the program.
To develop the HPL program, the researchers first analyzed lung adenocarcinoma tissue slides from the Cancer Genome Atlas. Adenocarcinoma was chosen for the test model because the disease is known for characteristic features. As an example, they note that its tumor cells tend to group in so-called acinar, or saclike patterns and spread predictably along the surface lining of lung cells.
From their analysis of the slides, whose visual images were digitally scanned and broken into 432,231 small quadrants or tiles, researchers found 46 key characteristics, what they term histomorphological phenotype clusters, from both normal and diseased tissue, a subset of which were statistically linked to either cancer’s early return or to long-term survival. The findings were then confirmed by further and separate testing on tissue images from 276 men and women who were treated for adenocarcinoma at NYU Langone from 2006 to 2021.
Researchers say their goal is to use the HPL algorithm to assign to each patient a score between 0 and 1 that reflects their statistical chance of survival and tumor recurrence for up to five years. Because the program is self-learning, they stress HPL will become increasingly more accurate as more data is added over time. To build public trust, researchers have posted their programming code online and have plans to make the new HPL tool freely available upon completion of further testing.
Characteristics linked to tumors recurring included high tile percentages of dead cancer cells and tumor-fighting immune cells called lymphocytes, and the dense clustering of tumor cells in the outer linings of the lungs. Features tied to increased likelihood for survival were high percentages of unchanged or preserved lung sac tissue, and lack of or mild presence of inflammatory cells.
Tsirigos says the team next plans to look at developing HPL-like programs for other cancers, such as breast, ovarian, and colorectal, that are similarly based on distinctive and key morphological features and additional molecular data. The team also has plans to expand and improve the accuracy of the current adenocarcinoma HPL program by including other data from hospital electronic health records about other illnesses and diseases, or even income and home ZIP code.
Funding support for the new study was provided by National Institutes of Health grant P30CA016087, United Kingdom Research Council grants Ep/R018634/1 and BB/V016067/1, and European Union Horizon 2020 grant no. 101016851.
Besides Tsirigos and Coudray, other NYU Langone researchers involved in this study are Anna Yeaton, Bojing Liu, Hortense Le, Luis Chiriboga, Afreen Karimkhan, Navneet Natula, Christopher Park, Harvey Pass, and Andre Moreira. Study co-lead investigator Adalberto Claudio Quiros, study co-investigators Xinyu Yang and John Le Quesne, and study co-senior investigator Ke Yuan are all at the University of Glasgow, UK. Study co-investigator David Moore is at the University College London, UK. More