Computers Math

In a perfect world, what you see is what you get. If this were the case, the job of artificial intelligence systems would be refreshingly straightforward.
Take collision avoidance systems in self-driving cars. If visual input to on-board cameras could be trusted entirely, an AI system could directly map that input to an appropriate action — steer right, steer left, or continue straight — to avoid hitting a pedestrian that its cameras see in the road.
But what if there’s a glitch in the cameras that slightly shifts an image by a few pixels? If the car blindly trusted so-called “adversarial inputs,” it might take unnecessary and potentially dangerous action.
A new deep-learning algorithm developed by MIT researchers is designed to help machines navigate in the real, imperfect world, by building a healthy “skepticism” of the measurements and inputs they receive.
The team combined a reinforcement-learning algorithm with a deep neural network, both used separately to train computers in playing video games like Go and chess, to build an approach they call CARRL, for Certified Adversarial Robustness for Deep Reinforcement Learning.
The researchers tested the approach in several scenarios, including a simulated collision-avoidance test and the video game Pong, and found that CARRL performed better — avoiding collisions and winning more Pong games — over standard machine-learning techniques, even in the face of uncertain, adversarial inputs.

advertisement

“You often think of an adversary being someone who’s hacking your computer, but it could also just be that your sensors are not great, or your measurements aren’t perfect, which is often the case,” says Michael Everett, a postdoc in MIT’s Department of Aeronautics and Astronautics (AeroAstro). “Our approach helps to account for that imperfection and make a safe decision. In any safety-critical domain, this is an important approach to be thinking about.”
Everett is the lead author of a study outlining the new approach, which appears in IEEE’s Transactions on Neural Networks and Learning Systems. The study originated from MIT PhD student Björn Lütjens’ master’s thesis and was advised by MIT AeroAstro Professor Jonathan How.
Possible realities
To make AI systems robust against adversarial inputs, researchers have tried implementing defenses for supervised learning. Traditionally, a neural network is trained to associate specific labels or actions with given inputs. For instance, a neural network that is fed thousands of images labeled as cats, along with images labeled as houses and hot dogs, should correctly label a new image as a cat.
In robust AI systems, the same supervised-learning techniques could be tested with many slightly altered versions of the image. If the network lands on the same label — cat — for every image, there’s a good chance that, altered or not, the image is indeed of a cat, and the network is robust to any adversarial influence.

advertisement

But running through every possible image alteration is computationally exhaustive and difficult to apply successfully to time-sensitive tasks such as collision avoidance. Furthermore, existing methods also don’t identify what label to use, or what action to take, if the network is less robust and labels some altered cat images as a house or a hotdog.
“In order to use neural networks in safety-critical scenarios, we had to find out how to take real-time decisions based on worst-case assumptions on these possible realities,” Lütjens says.
The best reward
The team instead looked to build on reinforcement learning, another form of machine learning that does not require associating labeled inputs with outputs, but rather aims to reinforce certain actions in response to certain inputs, based on a resulting reward. This approach is typically used to train computers to play and win games such as chess and Go.
Reinforcement learning has mostly been applied to situations where inputs are assumed to be true. Everett and his colleagues say they are the first to bring “certifiable robustness” to uncertain, adversarial inputs in reinforcement learning.
Their approach, CARRL, uses an existing deep-reinforcement-learning algorithm to train a deep Q-network, or DQN — a neural network with multiple layers that ultimately associates an input with a Q value, or level of reward.
The approach takes an input, such as an image with a single dot, and considers an adversarial influence, or a region around the dot where it actually might be instead. Every possible position of the dot within this region is fed through a DQN to find an associated action that would result in the most optimal worst-case reward, based on a technique developed by recent MIT graduate student Tsui-Wei “Lily” Weng PhD ’20.
An adversarial world
In tests with the video game Pong, in which two players operate paddles on either side of a screen to pass a ball back and forth, the researchers introduced an “adversary” that pulled the ball slightly further down than it actually was. They found that CARRL won more games than standard techniques, as the adversary’s influence grew.
“If we know that a measurement shouldn’t be trusted exactly, and the ball could be anywhere within a certain region, then our approach tells the computer that it should put the paddle in the middle of that region, to make sure we hit the ball even in the worst-case deviation,” Everett says.
The method was similarly robust in tests of collision avoidance, where the team simulated a blue and an orange agent attempting to switch positions without colliding. As the team perturbed the orange agent’s observation of the blue agent’s position, CARRL steered the orange agent around the other agent, taking a wider berth as the adversary grew stronger, and the blue agent’s position became more uncertain.
There did come a point when CARRL became too conservative, causing the orange agent to assume the other agent could be anywhere in its vicinity, and in response completely avoid its destination. This extreme conservatism is useful, Everett says, because researchers can then use it as a limit to tune the algorithm’s robustness. For instance, the algorithm might consider a smaller deviation, or region of uncertainty, that would still allow an agent to achieve a high reward and reach its destination.
In addition to overcoming imperfect sensors, Everett says CARRL may be a start to helping robots safely handle unpredictable interactions in the real world.
“People can be adversarial, like getting in front of a robot to block its sensors, or interacting with them, not necessarily with the best intentions,” Everett says. “How can a robot think of all the things people might try to do, and try to avoid them? What sort of adversarial models do we want to defend against? That’s something we’re thinking about how to do.”
This research was supported, in part, by Ford Motor Company as part of the Ford-MIT Alliance. More

Scientists have taken a major step forward in harnessing machine learning to accelerate the design for better batteries: Instead of using it just to speed up scientific analysis by looking for patterns in data, as researchers generally do, they combined it with knowledge gained from experiments and equations guided by physics to discover and explain a process that shortens the lifetimes of fast-charging lithium-ion batteries.
It was the first time this approach, known as “scientific machine learning,” has been applied to battery cycling, said Will Chueh, an associate professor at Stanford University and investigator with the Department of Energy’s SLAC National Accelerator Laboratory who led the study. He said the results overturn long-held assumptions about how lithium-ion batteries charge and discharge and give researchers a new set of rules for engineering longer-lasting batteries.
The research, reported today in Nature Materials, is the latest result from a collaboration between Stanford, SLAC, the Massachusetts Institute of Technology and Toyota Research Institute (TRI). The goal is to bring together foundational research and industry know-how to develop a long-lived electric vehicle battery that can be charged in 10 minutes.
“Battery technology is important for any type of electric powertrain,” said Patrick Herring, senior research scientist for Toyota Research Institute. “By understanding the fundamental reactions that occur within the battery we can extend its life, enable faster charging and ultimately design better battery materials. We look forward to building on this work through future experiments to achieve lower-cost, better-performing batteries.”
A trio of advances
The new study builds on two previous advances where the group used more conventional forms of machine learning to dramatically accelerate both battery testing and the process of winnowing down many possible charging methods to find the ones that work best. While these studies allowed researchers to make much faster progress — reducing the time needed to determine battery lifetimes by 98%, for instance — they didn’t reveal the underlying physics or chemistry that made some batteries last longer than others, as the latest study did.

advertisement

Combining all three approaches could potentially slash the time needed to bring a new battery technology from the lab bench to the consumer by as much as two-thirds, Chueh said.
“In this case, we are teaching the machine how to learn the physics of a new type of failure mechanism that could help us design better and safer fast-charging batteries,” Chueh said. “Fast charging is incredibly stressful and damaging to batteries, and solving this problem is key to expanding the nation’s fleet of electric vehicles as part of the overall strategy for fighting climate change.”
The new combined approach can also be applied to developing the grid-scale battery systems needed for a greater deployment of wind and solar electricity, which will become even more urgent as the nation pursues recently announced Biden Administration goals of eliminating fossil fuels from electric power generation by 2035 and achieving net-zero carbon emissions by 2050.
Zooming in for closeups
The new study zoomed in on battery electrodes, which are made of nano-sized grains glommed together into particles. Lithium ions slosh back and forth between the cathode and anode during charging and discharging, seeping into the particles and flowing back out again. This constant back-and-forth makes particles swell, shrink and crack, gradually decreasing their ability to store charge, and fast charging just makes things worse.

advertisement

To look at this process in more detail, the team observed the behavior of cathode particles made of nickel, manganese and cobalt, a combination known as NMC that’s one of the most widely used materials in electric vehicle batteries. These particles absorb lithium ions when the battery discharges and release them when it charges.
Stanford postdoctoral researchers Stephen Dongmin Kang and Jungjin Park used X-rays from SLAC’s Stanford Synchrotron Radiation Lightsource to get an overall look at particles that were undergoing fast charging. Then they took particles to Lawrence Berkeley National Laboratory’s Advanced Light Source to be examined with scanning X-ray transmission microscopy, which homes in on individual particles.
The data from those experiments, along with information from mathematical models of fast charging and equations that describe the chemistry and physics of the process, were incorporated into scientific machine learning algorithms.
“Rather than having the computer directly figure out the model by simply feeding it data, as we did in the two previous studies, we taught the computer how to choose or learn the right equations, and thus the right physics,” said Kang, who performed the modeling with MIT graduate student Hongbo Zhao, working with chemical engineering professor Martin Bazant.
The rich-get-richer effect
Until now, scientists had assumed that the differences between particles were insignificant and that their ability to store and release ions was limited by how fast lithium could move inside the particles, Kang said. In this way of seeing things, lithium ions flow in and out of all the particles at the same time and at roughly the same speed.
But the new approach revealed that the particles themselves control how fast lithium ions move out of cathode particles when a battery charges, he said. Some particles immediately release a lot of their ions while others release very few or none at all. And the quick-to-release particles go on releasing ions at a faster rate than their neighbors – a positive feedback, or “rich get richer,” effect that had not been identified before.
“We now have a picture — literally a movie — of how lithium moves around inside the battery, and it’s very different than scientists and engineers thought it was,” Kang said. “This uneven charging and discharging puts more stress on the electrodes and decreases their working lifetimes. Understanding this process on a fundamental level is an important step toward solving the fast charging problem.”
The scientists say their new method has potential for improving the cost, storage capacity, durability and other important properties of batteries for a wide range of applications, from electric vehicles to laptops to large-scale storage of renewable energy on the grid.
“It was just two years ago that the 2019 Nobel Prize in chemistry was awarded for the development of rechargeable lithium-ion batteries, which dates back to the 1970s,” Chueh said. “So I am encouraged that there’s still so much to learn about how to make batteries better.”
This research was funded by Toyota Research Institute. The Stanford Synchrotron Radiation Lightsource and Advanced Light Source are DOE Office of Science user facilities, and work there was supported by the DOE Office of Science and the DOE Advanced Battery Materials Research Program. More

Virtual reality avatar-based coaching shows promise to increase access to and extend the reach of nutrition education programs to children at risk for obesity, according to a new study in the Journal of Nutrition Education and Behavior, published by Elsevier.
Researchers introduced 15 adult-child dyads to a virtual avatar-based coaching program that incorporated age-specific information on growth; physical, social, and emotional development; healthy lifestyles; common nutrition concerns; and interview questions around eating behaviors and food resources and counseling.
“We developed a virtual reality avatar computer program as a way to get kids engaged in learning about nutrition education. The goal was to make this a program that could work to prevent childhood obesity,” said lead investigator Jared T. McGuirt, PhD, MPH, Department of Nutrition, University of North Carolina Greensboro, Greensboro, NC, USA. “We were primarily interested in how kids and parents reacted to this program — particularly lower income kids and parents who may not have been able to access this kind of experience in the past.”
A key finding in the study was the avatar’s ability to spark dialogue between the children and adults around dietary habits and behavior. All children and adults reported liking the program and planned to use it in the future, as they found it fun, informational, and motivating. The personalized social aspect of the avatar experience was appealing, as participants thought the avatar would reinforce guidance and provide support while acting as a cue to change health behaviors.
Looking to future implementation of this program in the public health field, Dr. McGuirt noted: “We feel we have a good start. The program was designed so others can build on it, and hopefully advance this technique into community nutrition education programs.

Story Source:
Materials provided by Elsevier. Note: Content may be edited for style and length. More

North Carolina State University engineers continue to improve the efficiency of a flexible device worn on the wrist that harvests heat energy from the human body to monitor health.
In a paper published in npj Flexible Electronics, the NC State researchers report significant enhancements in preventing heat leakage in the flexible body heat harvester they first reported in 2017 and updated in 2020. The harvesters use heat energy from the human body to power wearable technologies — think of smart watches that measure your heart rate, blood oxygen, glucose and other health parameters — that never need to have their batteries recharged. The technology relies on the same principles governing rigid thermoelectric harvesters that convert heat to electrical energy.
Flexible harvesters that conform to the human body are highly desired for use with wearable technologies. Mehmet Ozturk, an NC State professor of electrical and computer engineering and the corresponding author of the paper, mentioned superior skin contact with flexible devices, as well as the ergonomic and comfort considerations to the device wearer, as the core reasons behind building flexible thermoelectric generators, or TEGs.
The performance and efficiency of flexible harvesters, however, historically trail well behind rigid devices, which have been superior in their ability to convert body heat into usable energy.
The NC State proof-of-concept TEG originally reported in 2017 employed semiconductor elements that were connected electrically in series using liquid-metal interconnects made of EGaIn — a non-toxic alloy of gallium and indium. EGaIn provided both metal-like electrical conductivity and stretchability. The entire device was embedded in a stretchable silicone elastomer.
The upgraded device reported in 2020 employed the same architecture but significantly improved the thermal engineering of the previous version, while increasing the density of the semiconductor elements responsible for converting heat into electricity. One of the improvements was a high thermal conductivity silicone elastomer — essentially a type of rubber — that encapsulated the EGaIn interconnects.
The newest iteration adds aerogel flakes to the silicone elastomer to reduce the elastomer’s thermal conductivity. Experimental results showed that this innovation reduced the heat leakage through the elastomer by half.
“The addition of aerogel stops the heat from leaking between the device’s thermoelectric ‘legs,'” Ozturk said. “The higher the heat leakage, the lower the temperature that develops across the device, which translates to lower output power.
“The flexible device reported in this paper is performing an order of magnitude better than the device we reported in 2017 and continues to approach the performance of rigid devices,” Ozturk added.
Ozturk said that one of the strengths of the NC State-patented technology is that it employs the very same semiconductor elements used in rigid devices perfected after decades of research. The approach also provides a low-cost opportunity to existing rigid thermoelectric module manufacturers to enter the flexible thermoelectric market.
He added that his lab will continue to focus on improving the efficiency of these flexible devices.

Story Source:
Materials provided by North Carolina State University. Original written by Mick Kulikowski. Note: Content may be edited for style and length. More

Public grants to build rural broadband networks may not be sufficient to close the digital divide, new Cornell University research finds.
High operations and maintenance costs and low population density in some rural areas result in prohibitively high service fees — even for a subscriber-owned cooperative structured to prioritize member needs over profits, the analysis found.
Decades ago, cooperatives were key to the expansion of electric and telephone service to underserved rural areas, spurred by New Deal legislation providing low-interest government grants and loans. Public funding for rural broadband access should similarly consider its critical role supporting economic development, health care and education, said Todd Schmit, associate professor in the Charles H. Dyson School of Applied Economics and Management.
“The New Deal of broadband has to incorporate more than building the systems,” Schmit said. “We have to think more comprehensively about the importance of getting equal access to these technologies.”
Schmit is the co-author with Roberta Severson, an extension associate in Dyson, of “Exploring the Feasibility of Rural Broadband Cooperatives in the United States: The New New Deal?” The research was published Feb. 13 in Telecommunications Policy.
More than 90% of Americans had broadband access in 2015, according to the study, but the total in rural areas was below 70%. Federal programs have sought to help close that gap, including a $20.4 billion Federal Communications Commission initiative announced last year to subsidize network construction in underserved areas.

advertisement

Schmit and Severson studied the feasibility of establishing a rural broadband cooperative to improve access in Franklin County in northern New York state, which received funding for a feasibility study from the U.S. Department of Agriculture’s Rural Business Development Program.
The researchers partnered with Slic Network Solutions, a local internet service provider, to develop estimates of market prices, the cost to build a fiber-to-the-home network, operations and maintenance costs, and the potential subscriber base — about 1,600 residents — and model a cooperative that would break even over a 10-year cycle.
Federal and state grants and member investment would cover almost the entire estimated $8 million construction cost, so that wasn’t a significant factor in the analysis, the researchers said.
But even with those subsidies, the study determined the co-op would need to charge $231 per month for its high-speed service option — 131% above market rates. At that price, it’s unlikely 40% of year-round residents would opt for high-speed broadband as the model had assumed, casting further doubt on its feasibility.
The $231 fee included a surcharge to subsidize a lower-speed service option costing no more than $60 — a restriction the construction grants imposed to ensure affordability. Without that restriction, the high-speed price would drop to $175 and the low-speed climb to $105.

advertisement

“In short,” the authors wrote, “grants covering investment and capital construction alone do not solve the rural broadband problem, at least in our study area.”
As an alternative — though not one available in Franklin County — Schmit and Severson examined the possibility of an existing rural electric or telecommunications co-op expanding into broadband. They would gain efficiencies from already operating infrastructure such as the poles that would carry fiber lines. In that scenario, the high-speed price improved to $144 a month — still 44% above market rates.
“These systems are very costly to operate and maintain,” Schmit said, “particularly in areas like we looked at that are very low density.”
The feasibility improves with growth in a coverage area’s density and “take rate,” or percentage of potential subscribers signing up at different speeds, according to the analysis. But in Franklin County, the researchers determined a startup co-op would need 14 potential subscribers per mile to break even over 10 years — more than twice the study area’s actual density.
To better serve such areas, Schmit and Severson said, policymakers should explore eliminating property taxes on broadband infrastructure and payments to rent space on poles owned by regulated utilities, which respectively accounted for 16% and 18% of the proposed co-op’s annual expenses. Those measures reduced an expanding rural utility co-op’s high-speed fee to 25% above market rates, a level members might be willing to pay, the authors said.
“Consideration of the public benefits of broadband access arguably needs to be added to the equation,” they wrote. “The case was made for electricity and telephone services in the 1930s and similar arguments would seem to hold for this technology today.” More

Egg cells are by far the largest cells produced by most organisms. In humans, they are several times larger than a typical body cell and about 10,000 times larger than sperm cells.
There’s a reason why egg cells, or oocytes, are so big: They need to accumulate enough nutrients to support a growing embryo after fertilization, plus mitochondria to power all of that growth. However, biologists don’t yet understand the full picture of how egg cells become so large.
A new study in fruit flies, by a team of MIT biologists and mathematicians, reveals that the process through which the oocyte grows significantly and rapidly before fertilization relies on physical phenomena analogous to the exchange of gases between balloons of different sizes. Specifically, the researchers showed that “nurse cells” surrounding the much larger oocyte dump their contents into the larger cell, just as air flows from a smaller balloon into a larger one when they are connected by small tubes in an experimental setup.
“The study shows how physics and biology come together, and how nature can use physical processes to create this robust mechanism,” says Jörn Dunkel, an MIT associate professor of physical applied mathematics. “If you want to develop as an embryo, one of the goals is to make things very reproducible, and physics provides a very robust way of achieving certain transport processes.”
Dunkel and Adam Martin, an MIT associate professor of biology, are the senior authors of the paper, which appears this week in the Proceedings of the National Academy of Sciences. The study’s lead authors are postdoc Jasmin Imran Alsous and graduate student Nicolas Romeo. Jonathan Jackson, a Harvard University graduate student, and Frank Mason, a research assistant professor at Vanderbilt University School of Medicine, are also authors of the paper.
A physical process
In female fruit flies, eggs develop within cell clusters known as cysts. An immature oocyte undergoes four cycles of cell division to produce one egg cell and 15 nurse cells. However, the cell separation is incomplete, and each cell remains connected to the others by narrow channels that act as valves that allow material to pass between cells.

advertisement

Members of Martin’s lab began studying this process because of their longstanding interest in myosin, a class of proteins that can act as motors and help muscle cells contract. Imran Alsous performed high-resolution, live imaging of egg formation in fruit flies and found that myosin does indeed play a role, but only in the second phase of the transport process. During the earliest phase, the researchers were puzzled to see that the cells did not appear to be increasing their contractility at all, suggesting that a mechanism other than “squeezing” was initiating the transport.
“The two phases are strikingly obvious,” Martin says. “After we saw this, we were mystified, because there’s really not a change in myosin associated with the onset of this process, which is what we were expecting to see.”
Martin and his lab then joined forces with Dunkel, who studies the physics of soft surfaces and flowing matter. Dunkel and Romeo wondered if the cells might be behaving the same way that balloons of different sizes behave when they are connected. While one might expect that the larger balloon would leak air to the smaller until they are the same size, what actually happens is that air flows from the smaller to the larger.
This happens because the smaller balloon, which has greater curvature, experiences more surface tension, and therefore higher pressure, than the larger balloon. Air is therefore forced out of the smaller balloon and into the larger one. “It’s counterintuitive, but it’s a very robust process,” Dunkel says.
Adapting mathematical equations that had already been derived to explain this “two-balloon effect,” the researchers came up with a model that describes how cell contents are transferred from the 15 small nurse cells to the large oocyte, based on their sizes and their connections to each other. The nurse cells in the layer closest to the oocyte transfer their contents first, followed by the cells in more distant layers.

advertisement

“After I spent some time building a more complicated model to explain the 16-cell problem, we realized that the simulation of the simpler 16-balloon system looked very much like the 16-cell network. It is surprising to see that such counterintuitive but mathematically simple ideas describe the process so well,” Romeo says.
The first phase of nurse cell dumping appears to coincide with when the channels connecting the cells become large enough for cytoplasm to move through them. Once the nurse cells shrink to about 25 percent of their original size, leaving them only slightly larger than their nuclei, the second phase of the process is triggered and myosin contractions force the remaining contents of the nurse cells into the egg cell.
“In the first part of the process, there’s very little squeezing going on, and the cells just shrink uniformly. Then this second process kicks in toward the end where you start to get more active squeezing, or peristalsis-like deformations of the cell, that complete the dumping process,” Martin says.
Cell cooperation
The findings demonstrate how cells can coordinate their behavior, using both biological and physical mechanisms, to bring about tissue-level behavior, Imran Alsous says.
“Here, you have several nurse cells whose job it is to nurse the future egg cell, and to do so, these cells appear to transport their contents in a coordinated and directional manner to the oocyte,” she says.
Oocyte and early embryonic development in fruit flies and other invertebrates bears some similarities to those of mammals, but it’s unknown if the same mechanism of egg cell growth might be seen in humans or other mammals, the researchers say.
“There’s evidence in mice that the oocyte develops as a cyst with other interconnected cells, and that there is some transport between them, but we don’t know if the mechanisms that we’re seeing here operate in mammals,” Martin says.
The researchers are now studying what triggers the second, myosin-powered phase of the dumping process to start. They are also investigating how changes to the original sizes of the nurse cells might affect egg formation.
The research was funded by the National Institute of General Medical Sciences, a Complex Systems Scholar Award from the James S. McDonnell Foundation, and the Robert E. Collins Distinguished Scholarship Fund. More

Professor Byoungwoo Kang develops a high-density cathode material through controlling local structures of the Li-rich layered materials.
Researchers in Korea have developed a high-capacity cathode material that can be stably charged and discharged for hundreds of cycles without using the expensive cobalt (Co) metal. The day is fast approaching when electric vehicles can drive long distances with Li- ion batteries.
Professor Byoungwoo Kang and Dr. Junghwa Lee of POSTECH’s Department of Materials Science and Engineering have successfully developed a high energy-density cathode material that can stably maintain charge and discharge for over 500 cycles without the expensive and toxic Co metal. The research team achieved this by controlling the local structure via developing the simple synthesis process for the Li-rich layered material that is attracting attention as the next-generation high-capacity cathode material. These research findings were published in ACS Energy Letters, a journal in the energy field of the American Chemistry Association.
The mileage and the charge-discharge cycle of an electric vehicle depend on the unique properties of the electrode material in the rechargeable Li-ion battery. Electricity is generated when lithium ions flow back and forth between the cathode and anode. In the case of Li-rich layered material, the number of cycles decreases sharply when large amount of lithium is extracted and inserted. In particular, when a large amount of lithium is extracted and an oxygen reaction occurs in a highly charged state, a structural collapse occurs rendering it impossible to maintain the charge-discharge properties or the high-energy density for long-term cycles. This deterioration of the cycle property has hampered commercialization.
The research team had previously revealed that the homogeneous distribution of atoms between the transition metal layer and the lithium layer of the Li-rich layered material can be an important factor in the activation of electrochemical reaction and cycle property in Li-rich layered materials. The team then conducted a subsequent research to control the synthesis conditions for increasing the degree of the atoms’ distribution in the structure. Using the previously published solid-state reaction, the team developed a new, simple and efficient process that can produce a cathode material that has an optimized atomic distribution.
As a result, it was confirmed that the synthesized Li-rich layered material has an optimized local structure in terms of electrochemical activity and cycle property, enabling a large amount of lithium to be used reversibly. It was also confirmed that the oxygen redox reaction was also stably and reversibly driven for several hundred cycles.
Under these optimized conditions, the Co-free Li-rich layered material synthesized displayed 180% higher reversible energy at 1,100Wh/kg than the conventionally commercialized high nickel layered material (ex. LiNi0.8Mn0.1Co0.1O2) with energy density of 600Wh/kg. In particular, even if a large amount of lithium is removed, a stable structure was maintained, enabling about 95% capacity for 100 cycles. In addition, by maintaining 83% for 500 cycles, a breakthrough performance that can maintain stable high energy for hundreds of cycles is anticipated.
“The significance of these research findings is that the cycle property, which is one of the important issues in the next-generation high-capacity Li-rich layered materials, have been dramatically improved through relatively simple process changes,” explained Professor Byoungwoo Kang of POSTECH. “This is noteworthy in that we have moved a step closer to commercializing the next generation Li-rich layered materials.”

Story Source:
Materials provided by Pohang University of Science & Technology (POSTECH). Note: Content may be edited for style and length. More

A new data-driven study from Texas A&M University casts serious doubt on the stereotype that male students perform better than female students in science — specifically, physics.
A team of researchers in the Department of Physics and Astronomy analyzed both the midterm exam scores and final grades of more than 10,000 Texas A&M students enrolled in four introductory physics courses across more than a decade, finding no evidence that male students consistently outperform female students in these courses.
The work was led by Texas A&M physicist and Presidential Professor for Teaching Excellence Tatiana Erukhimova.
With help of nearly two dozen departmental colleagues, the Texas A&M team built a database reflecting the complete introductory physics educational spectrum: the calculus-based course sequence primarily taken by engineering and physics majors as well as the algebra-based course sequence typically taken by life sciences and premed majors. Their final analysis shows that exam performance and final letter grades are largely independent of student gender — results which Erukhimova says show promise in ending gender stereotypes that negatively impact so many female students in STEM.
“There is no consistent trend on male students outperforming female students,” Erukhimova said. “Our study also provides new knowledge regarding whether statistically significant differences based on gender occurred on each exam for four introductory physics courses as the semesters were progressing — an area that has not previously been studied, at least not for such a large data set and over a long period of time.”
When differences in final letter grades for a course were observed, there were no persistent differences across that course’s exams, she said. Conversely, when researchers found differences on exams within a course, they observed no differences for final letter grades in that course. In algebra-based mechanics, they found that female students outperformed male students by a small but statistically significant margin.

advertisement

Their findings were published recently in the American Physical Society journal Physical Review Physics Education Research.
Prior to the team’s study and others similar to it, Erukhimova says it has been an open question as to whether significant differences between male and female students could show up on particular exams but remain slight enough so as not to affect final course grades. For the past 25 years, the physics education profession has relied on inventory tests — optional surveys intended to assess conceptual understanding and retention of key physics concepts — to answer that question, effectively substantiating the argument for gender differences in student performance by default because men tend to score higher on them.
“In the field of physics education research, the majority of existing studies report a persistent gender gap with males performing significantly better than females on introductory mechanics concept inventory assessments, such as the Force Concept Inventory,” Erukhimova said. “The results of prior studies on the gendered differences in student performance based on course grades and examinations are less consistent. While a number of studies indicate that male students outperform female students on the exams and course grades, other groups found no significant gendered difference in student performance.”
The team applied multiple statistical analyses to the course-level data they collected to study whether there were performance differences based on student gender. To see how their findings aligned with student perceptions, they also took a snapshot of the students’ feelings about course performance, inclusion and contributions using a short anonymous questionnaire distributed to 1,600 students in fall 2019.
“Responses indicated that female students had lower perception of their performance than their male classmates,” Erukhimova said. “The only class where female students perceived their performance as equal to their male classmates was algebra-based mechanics, in which females typically outperform males. Additionally, we found that although male and female students may feel differently regarding their performance and in-class contributions, they feel equally included in class.”
Although the team’s study represents clear progress to Erukhimova, she acknowledges it comes with its own limitations — the most significant being that it relies solely on course-level data collected from faculty and does not analyze the possible impact of non-academic factors on student performance. In the future, she says the team would like to connect as much of their data set as possible to university-level records to see how prior preparation, such as SAT scores, affects these results.
“We believe that all students should have equal opportunities and chances for success in physics,” Erukhimova said. “The results of this work may help with fighting the gender stereotype threat that negatively impacts so many female students. By contributing to the body of knowledge about how gender relates to student performance, we hope that our work, which would not have been possible without our colleagues’ data, can be another step in dismantling the preconceived notion of a societal bias based on gender in physics.”
The team’s research was funded in part by the College of Science Diversity and Equity Small Grants Program.

Story Source:
Materials provided by Texas A&M University. Original written by Shana Hutchins. Note: Content may be edited for style and length. More