Computers Math

Using computer-assisted neural networks, Researchers at the Karlsruhe Institute of Technology (KIT) and the University of Duisburg-Essen have been able to accurately identify affective states from the body language of tennis players during games. For the first time, they trained a model based on artificial intelligence (AI) with data from actual games. Their study, published in the journal Knowledge-Based Systems, demonstrates that AI can assess body language and emotions with accuracy similar to that of humans. However, it also points to ethical concerns.
For their study, “Recognizing affective states from the expressive behavior of tennis players using convolutional neural networks,” sports sciences, software development and computer science researchers from KIT and the University of Duisburg-Essen developed a special AI model. They used pattern-recognition programs to analyze video of tennis players recorded during actual games.
Success Rate of 68.9 Percent
“Our model can identify affective states with an accuracy of up to 68.9 percent, which is comparable and sometimes even superior to assessments made by both human observers and earlier automated methods,” said Professor Darko Jekauc of KIT’s Institute of Sports and Sports Science.
An important and unique feature of the study is the project team’s use of real-life scenes instead of simulated or contrived situations to train their AI system. The researchers recorded video sequences of 15 tennis players in a specific setting, focusing on the body language displayed when a point was won or lost. The videos showed players with cues including lowered head, arms raised in exultation, hanging racket, or differences in walking speed; these cues could be used to identify the players’ affective states.
After being fed with this data, the AI learned to associate the body language signals with different affective reactions and to determine whether a point had been won (positive body language) or lost (negative body language). “Training in natural contexts is a significant advance for the identification of real emotional states, and it makes predictions possible in real scenarios,” said Jekauc.
Humans and Machines Recognize Negative Emotions Better Than Positive Ones
Not only does the research show that AI algorithms may be able to surpass human observers in their ability to identify emotions in the future, it also revealed a further interesting aspect: both humans and AI are better at recognizing negative emotions. “The reason could be that negative emotions are easier to identify because they’re expressed in more obvious ways,” said Jekauc. “Psychological theories suggest that people are evolutionarily better adapted to perceive negative emotional expressions, for example because defusing conflict situations quickly is essential to social cohesion.”

Ethical Aspects Need Clarification Before Use
The study envisions a number of sports applications for reliable emotion recognition, such as improving training methods, team dynamics and performance, and preventing burnout. Other fields, including healthcare, education, customer service and automotive safety, could also benefit from reliable early detection of emotional states.
“Although this technology offers the prospect of significant benefits, the potential risks associated with it also have to be taken into account, especially those relating to privacy and misuse of data,” Jekauc said. “Our study adhered strictly to existing ethical guidelines and data protection regulations. And with a view to future applications of such technology in practice, it will be essential to clarify ethical and legal issues ahead of time.” More

Porous covalent organic frameworks (COFs) are a class of highly ordered, porous materials consisting of organic molecules that are linked by covalent bonds to form a network. They enable the construction of functional materials with molecular precision. Similar to metal organic frameworks (MOFs), which were discovered around 25 years ago and have already reached market maturity, COFs possess highly promising structural, optical and electronic properties for numerous applications, for example in gas and liquid storage, catalysis, sensor technology and energy applications.

Previous research on COFs has generally focussed on the construction of rigid frameworks with static material properties. Dr Florian Auras and his team at the Chair of Molecular Functional Materials at TUD have now developed a design strategy for dynamic two-dimensional COFs that can open and close their pores in a controlled manner, similar to a sponge. “The main aim of the study was to equip these frameworks, which are normally very precisely ordered but rigid, with exactly the right degree of flexibility so that their structure can be switched from compact to porous. By adding solvent to the molecular sponge, we can now temporarily and reversibly change the local geometry as well as optical properties such as colour or fluorescence,” says Florian Auras, explaining his research approach.
The ability to switch the structural and optoelectronic properties of the materials back and forth in a targeted manner makes the materials particularly interesting for future applications in electronics and information technology. “Our research results form the basis for our further research into stimuli-responsive polymers, particularly with the aim of realising switchable quantum states. When working on COFs, I am always fascinated by how precisely their properties can be manipulated by controlling the molecular structure,” adds Auras. More

The United States is experiencing a housing shortage. At least, that is the case according to common belief — and is even the basis for national policy, as the Biden administration has stated plans to address the housing supply shortfall.
But new research from the University of Kansas finds that most of the nation’s markets have ample housing in total, but nearly all lack enough units affordable to very low-income households.
Kirk McClure, professor of public affairs & administration emeritus at KU, and Alex Schwartz of The New School co-wrote a study published in the journal Housing Policy Debate. They examined U.S. Census Bureau data from 2000 to 2020 to compare the number of households formed to the number of housing units added to determine if there were more households needing homes than units available.
The researchers found only four of the nation’s 381 metropolitan areas experienced a housing shortage in the study time frame, as did only 19 of the country’s 526 “micropolitan” areas — those with 10,000-50,000 residents.
The findings suggest that addressing housing prices and low incomes are more urgently needed to address housing affordability issues than simply building more homes, the authors wrote.
“There is a commonly held belief that the United States has a shortage of housing. This can be found in the popular and academic literature and from the housing industry,” McClure said. “But the data shows that the majority of American markets have adequate supplies of housing available. Unfortunately, not enough of it is affordable, especially for low-income and very low-income families and individuals.”
McClure and Schwartz also examined households in two categories: Very low income, defined as between 30% and 60% of area median family income, and extremely low income, with incomes below 30% of area median family income.

The numbers showed that from 2010 to 2020, household formation did exceed the number of homes available. However, there was a large surplus of housing produced in the previous decade. In fact, from 2000 to 2020, housing production exceeded the growth of households by 3.3 million units. The surplus from 2000 to 2010 more than offset the shortages from 2010 to 2020.
The numbers also showed that nearly all metropolitan areas have sufficient units for owner occupancy. But nearly all have shortages of rental units affordable to the very low-income renter households.
While the authors looked at housing markets across the nation, they also examined vacancy rates, or the difference between total and occupied units, to determine how many homes were available. National total vacancy rates were 9% in 2000 and 11.4% by 2010, which marked the end of the housing bubble and the Great Recession. By the end of 2020, the rate was 9.7%, with nearly 14 million vacant units.
“When looking at the number of housing units available, it becomes clear there is no overall shortage of housing units available. Of course, there are many factors that determine if a vacant is truly available; namely, if it is physically habitable and how much it costs to purchase or rent the unit,” McClure said. “There are also considerations over a family’s needs such as an adequate number of bedrooms or accessibility for individuals with disabilities, but the number of homes needed has not outpaced the number of homes available.”
Not all housing markets are alike, and while there could be shortages in some, others could contain a surplus of available housing units. The study considered markets in all core-based statistical areas as defined by the Census Bureau. Metropolitan areas saw a nationwide surplus of 2.7 million more units than households in the 20-year study period, while micropolitan areas had a more modest surplus of about 300,000 units.
Numbers of available housing units and people only tell part of the story. An individual family needs to be able to afford housing, whether they buy or rent. Shortages of any scale appear in the data only when considering renters, the authors wrote. McClure and Schwartz compared the number of available units in four submarkets of each core-based statistical area to the estimated number of units affordable to renters with incomes from 30% to 60% of the area median family income. Those rates are roughly equivalent to the federal poverty level and upper level of eligibility for various rental assistance programs. Only two metropolitan areas had shortages for very-low-income renters, and only two had surpluses available for extremely-low-income renters.
Helping people afford the housing stock that is available would be more cost effective than expanding new home construction in the hope that additional supply would bring prices down, the authors wrote. Several federal programs have proven successful in helping renters and moderate-income buyers afford housing that would otherwise be out of reach.
“Our nation’s affordability problems result more from low incomes confronting high housing prices rather than from housing shortages,” McClure said. “This condition suggests that we cannot build our way to housing affordability. We need to address price levels and income levels to help low-income households afford the housing that already exists, rather than increasing the supply in the hope that prices will subside.” More

Researchers at the University of Bonn have developed software that can simulate the growth of field crops. To do this, they fed thousands of photos from field experiments into a learning algorithm. This enabled the algorithm to learn how to visualize the future development of cultivated plants based on a single initial image. Using the images created during this process, parameters such as leaf area or yield can be estimated accurately. The results have been published in the journal Plant Methods.
Which plants should I combine in what ratio to achieve the greatest possible yield? And how will my crop develop if I use manure instead of artificial fertilizers? In the future, farmers should increasingly be able to count on computer support when answering such questions.
Researchers from the University of Bonn have now taken a crucial step forward on the path towards this goal: “We have developed software that uses drone photos to visualize the future development of the plants shown,” explains Lukas Drees from the Institute of Geodesy and Geoinformation at the University of Bonn. The early career researcher is an employee in the PhenoRob Cluster of Excellence. The large-scale project based at the University of Bonn intends to drive forward the intelligent digitalization of agriculture to help farming become more environmentally friendly, without causing harvest yields to suffer.
A virtual glimpse into the future to aid decision-making
The computer program now presented by Drees and his colleagues in the journal Plant Methods is an important building block. It should eventually make it possible to simulate certain decisions virtually — for instance, to assess how the use of pesticides or fertilizers will affect crop yield.
For this to work, the program must be fed with drone photos from field experiments. “We took thousands of images over one growth period,” explains the doctoral researcher. “In this way, for example, we documented the development of cauliflower crops under certain conditions.” The researchers then trained a learning algorithm using these images. Afterwards, based on a single aerial image of an early stage of growth, this algorithm was able to generate images showing the future development of the crop in a new, artificially created image. The whole process is very accurate as long as the crop conditions are similar to those present when the training photos were taken. Consequently, the software does not take into account the effect of a sudden cold snap or steady rain lasting several days. However, it should learn in the future how growth is affected by influences such as these — as well as an increased use of fertilizers, for example. This should enable it to predict the outcome of certain interventions by the farmer.
“In addition, we used a second AI software that can estimate various parameters from plant photos, such as crop yield,” says Drees. “This also works with the generated images. It is thus possible to estimate quite precisely the subsequent size of the cauliflower heads at a very early stage in the growth period.”
Focus on polycultures

One area the researchers are focusing on is the use of polycultures. This refers to the sowing of different species in one field — such as beans and wheat. As plants have different requirements, they compete less with each other in a polyculture of this kind compared to a monoculture, where just one species is grown. This boosts yield. In addition, some species — beans are a good example of this — can bind nitrogen from the air and use it as a natural fertilizer. The other species, in this case wheat, also benefits from this.
“Polycultures are also less susceptible to pests and other environmental influences,” explains Drees. “However, how well the whole thing works very much depends on the combined species and their mixing ratio.” When results from many different mixing experiments are fed into learning algorithms, it is possible to derive recommendations as to which plants are particularly compatible and in what ratio.
Plant growth simulations on the basis of learning algorithms are a relatively new development. Process-based models have mostly been used for this purpose up to now. These — metaphorically speaking — have a fundamental understanding of what nutrients and environmental conditions certain plants need during their growth in order to thrive. “Our software, however, makes its statements solely based on the experience they have collected using the training images,” stresses Drees.
Both approaches complement each other. If they were to be combined in an appropriate manner, it could significantly improve the quality of the forecasts. “This is also a point that we are investigating in our study,” says the doctoral researcher: “How can we use process- and image-based methods so they benefit from each other in the best possible way?” More

In a new paper, researchers from Carnegie Mellon University and Stevens Institute of Technology show a new way of thinking about the fair impacts of AI decisions.
They draw on a well-established tradition known as social welfare optimization, which aims to make decisions fairer by focusing on the overall benefits and harms to individuals. This method can be used to evaluate the industry standard assessment tools for AI fairness, which look at approval rates across protected groups.
“In assessing fairness, the AI community tries to ensure equitable treatment for groups that differ in economic level, race, ethnic background, gender, and other categories,” explained John Hooker, professor of operations research at the Tepper School of Business at Carnegie Mellon, who coauthored the study and presented the paper at the International Conference on the Integration of Constraint Programming, Artificial Intelligence, and Operations Research (CPAIOR) on May 29 in Uppsala, Sweden. The paper received the Best Paper Award.
Imagine a situation where an AI system decides who gets approved for a mortgage or who gets a job interview. Traditional fairness methods might only ensure that the same percentage of people from different groups get approved.
But what if being denied a mortgage has a much bigger negative impact on someone from a disadvantaged group than on someone from an advantaged group? By employing a social welfare optimization method, AI systems can make decisions that lead to better outcomes for everyone, especially for those in disadvantaged groups.
The study focuses on “alpha fairness,” a method of finding a balance between being fair and getting the most benefit for everyone. Alpha fairness can be adjusted to balance fairness and efficiency more or less, depending on the situation.
Hooker and his co-authors show how social welfare optimization can be used to compare different assessments for group fairness currently used in AI. By using this method, we can understand the benefits of applying different group fairness tools in different contexts. It also ties these group fairness assessment tools to the larger world of fairness-efficiency standards used in economics and engineering.

Derek Leben, associate teaching professor of business ethics at the Tepper School, and Violet Chen, assistant professor at Stevens Institute of Technology, who received her Ph.D. from the Tepper School, coauthored the study.
“Common group fairness criteria in AI typically compare statistical metrics of AI-supported decisions across different groups, ignoring the actual benefits or harms of being selected or rejected,” said Chen. “We propose a direct, welfare-centric approach to assess group fairness by optimizing decision social welfare. Our findings offer new perspectives on selecting and justifying group fairness criteria.”
“Our findings suggest that social welfare optimization can shed light on the intensely discussed question of how to achieve group fairness in AI,” Leben said.
The study is important for both AI system developers and policymakers. Developers can create more equitable and effective AI models by adopting a broader approach to fairness and understanding the limitations of fairness measures. It also highlights the importance of considering social justice in AI development, ensuring that technology promotes equity across diverse groups in society.
The paper is published in CPAIOR 2024 Proceedings. More

A team of researchers led by Philip Walther at the University of Vienna carried out a pioneering experiment where they measured the effect of the rotation of Earth on quantum entangled photons. The work, just published in Science Advances, represents a significant achievement that pushes the boundaries of rotation sensitivity in entanglement-based sensors, potentially setting the stage for further exploration at the intersection between quantum mechanics and general relativity.
Optical Sagnac interferometers are the most sensitive devices to rotations. They have been pivotal in our understanding of fundamental physics since the early years of the last century, contributing to establish Einstein’s special theory of relativity. Today, their unparalleled precision makes them the ultimate tool for measuring rotational speeds, limited only by the boundaries of classical physics.
Interferometers employing quantum entanglement have the potential to break those bounds. If two or more particles are entangled, only the overall state is known, while the state of the individual particle remains undetermined until measurement. This can be used to obtain more information per measurement than would be possible without it. However, the promised quantum leap in sensitivity has been hindered by the extremely delicate nature of entanglement. Here is where the Vienna experiment made the difference. They built a giant optical fiber Sagnac interferometer and kept the noise low and stable for several hours. This enabled the detection of enough high-quality entangled photon pairs such to outperform the rotation precision of previous quantum optical Sagnac interferometers by a thousand times.
In a Sagnac interferometer, two particles travelling in opposite directions of a rotating closed path reach the starting point at different times. With two entangled particles, it becomes spooky: they behave like a single particle testing both directions simultaneously while accumulating twice the time delay compared to the scenario where no entanglement is present. This unique property is known as super-resolution. In the actual experiment, two entangled photons were propagating inside a 2-kilometer-long optical fiber wounded onto a huge coil, realizing an interferometer with an effective area of more than 700 square meters.
A significant hurdle the researchers faced was isolating and extracting Earth’s steady rotation signal. “The core of the matter,” explains lead author Raffaele Silvestri, “lays in establishing a reference point for our measurement, where light remains unaffected by Earth’s rotational effect. Given our inability to halt Earth’s from spinning, we devised a workaround: splitting the optical fiber into two equal-length coils and connecting them via an optical switch.” By toggling the switch on and off the researchers could effectively cancel the rotation signal at will, which also allowed them to extend the stability of their large apparatus. “We have basically tricked the light into thinking it’s in a non-rotating universe,” says Silvestri.
The experiment, which was conducted as part of the research network TURIS hosted by the University of Vienna and the Austrian Academy of Sciences, has successfully observed the effect of the rotation of Earth on a maximally entangled two-photon state. This confirms the interaction between rotating reference systems and quantum entanglement, as described in Einstein’s special theory of relativity and quantum mechanics, with a thousand-fold precision improvement compared to previous experiments. “That represents a significant milestone since, a century after the first observation of Earth’s rotation with light, the entanglement of individual quanta of light has finally entered the same sensitivity regimes,” says Haocun Yu, who worked on this experiment as a Marie-Curie Postdoctoral Fellow. “I believe our result and methodology will set the ground to further improvements in the rotation sensitivity of entanglement-based sensors. This could open the way for future experiments testing the behavior of quantum entanglement through the curves of spacetime,” adds Philip Walther. More

Someday, you may want your home robot to carry a load of dirty clothes downstairs and deposit them in the washing machine in the far-left corner of the basement. The robot will need to combine your instructions with its visual observations to determine the steps it should take to complete this task.
For an AI agent, this is easier said than done. Current approaches often utilize multiple hand-crafted machine-learning models to tackle different parts of the task, which require a great deal of human effort and expertise to build. These methods, which use visual representations to directly make navigation decisions, demand massive amounts of visual data for training, which are often hard to come by.
To overcome these challenges, researchers from MIT and the MIT-IBM Watson AI Lab devised a navigation method that converts visual representations into pieces of language, which are then fed into one large language model that achieves all parts of the multistep navigation task.
Rather than encoding visual features from images of a robot’s surroundings as visual representations, which is computationally intensive, their method creates text captions that describe the robot’s point-of-view. A large language model uses the captions to predict the actions a robot should take to fulfill a user’s language-based instructions.
Because their method utilizes purely language-based representations, they can use a large language model to efficiently generate a huge amount of synthetic training data.
While this approach does not outperform techniques that use visual features, it performs well in situations that lack enough visual data for training. The researchers found that combining their language-based inputs with visual signals leads to better navigation performance.
“By purely using language as the perceptual representation, ours is a more straightforward approach. Since all the inputs can be encoded as language, we can generate a human-understandable trajectory,” says Bowen Pan, an electrical engineering and computer science (EECS) graduate student and lead author of a paper on this approach.

Pan’s co-authors include his advisor, Aude Oliva, director of strategic industry engagement at the MIT Schwarzman College of Computing, MIT director of the MIT-IBM Watson AI Lab, and a senior research scientist in the Computer Science and Artificial Intelligence Laboratory (CSAIL); Philip Isola, an associate professor of EECS and a member of CSAIL; senior author Yoon Kim, an assistant professor of EECS and a member of CSAIL; and others at the MIT-IBM Watson AI Lab and Dartmouth College. The research will be presented at the Conference of the North American Chapter of the Association for Computational Linguistics.
Solving a vision problem with language
Since large language models are the most powerful machine-learning models available, the researchers sought to incorporate them into the complex task known as vision-and-language navigation, Pan says.
But such models take text-based inputs and can’t process visual data from a robot’s camera. So, the team needed to find a way to use language instead.
Their technique utilizes a simple captioning model to obtain text descriptions of a robot’s visual observations. These captions are combined with language-based instructions and fed into a large language model, which decides what navigation step the robot should take next.
The large language model outputs a caption of the scene the robot should see after completing that step. This is used to update the trajectory history so the robot can keep track of where it has been.

The model repeats these processes to generate a trajectory that guides the robot to its goal, one step at a time.
To streamline the process, the researchers designed templates so observation information is presented to the model in a standard form — as a series of choices the robot can make based on its surroundings.
For instance, a caption might say “to your 30-degree left is a door with a potted plant beside it, to your back is a small office with a desk and a computer,” etc. The model chooses whether the robot should move toward the door or the office.
“One of the biggest challenges was figuring out how to encode this kind of information into language in a proper way to make the agent understand what the task is and how they should respond,” Pan says.
Advantages of language
When they tested this approach, while it could not outperform vision-based techniques, they found that it offered several advantages.
First, because text requires fewer computational resources to synthesize than complex image data, their method can be used to rapidly generate synthetic training data. In one test, they generated 10,000 synthetic trajectories based on 10 real-world, visual trajectories.
The technique can also bridge the gap that can prevent an agent trained with a simulated environment from performing well in the real world. This gap often occurs because computer-generated images can appear quite different from real-world scenes due to elements like lighting or color. But language that describes a synthetic versus a real image would be much harder to tell apart, Pan says.
Also, the representations their model uses are easier for a human to understand because they are written in natural language.
“If the agent fails to reach its goal, we can more easily determine where it failed and why it failed. Maybe the history information is not clear enough or the observation ignores some important details,” Pan says.
In addition, their method could be applied more easily to varied tasks and environments because it uses only one type of input. As long as data can be encoded as language, they can use the same model without making any modifications.
But one disadvantage is that their method naturally loses some information that would be captured by vision-based models, such as depth information.
However, the researchers were surprised to see that combining language-based representations with vision-based methods improves an agent’s ability to navigate.
“Maybe this means that language can capture some higher-level information than cannot be captured with pure vision features,” he says.
This is one area the researchers want to continue exploring. They also want to develop a navigation-oriented captioner that could boost the method’s performance. In addition, they want to probe the ability of large language models to exhibit spatial awareness and see how this could aid language-based navigation.
This research is funded, in part, by the MIT-IBM Watson AI Lab. More

Researchers from Tohoku University and Kyoto University have successfully developed a DNA-based molecular controller that autonomously directs the assembly and disassembly of molecular robots. This pioneering technology marks a significant step towards advanced autonomous molecular systems with potential applications in medicine and nanotechnology.
“Our newly developed molecular controller, composed of artificially designed DNA molecules and enzymes, coexists with molecular robots and controls them by outputting specific DNA molecules,” points out Shin-ichiro M. Nomura, an associate professor at Tohoku University’s Graduate School of Engineering and co-author of the study. “This allows the molecular robots to self-assemble and disassemble automatically, without the need for external manipulation.”
Such autonomous operation is a crucial advancement, as it enables the molecular robots to perform tasks in environments where external signals cannot reach.
In addition to Nomura, the research team included Ibuki Kawamata (an associate professor at Kyoto University’s Graduate School of Science), Kohei Nishiyama (a graduate student at Johannes Gutenberg University Mainz), and Akira Kakugo (a professor at Kyoto University’s Graduate School of Science).
Research on molecular robots, which are designed to aid in disease treatment and diagnosis by functioning both inside and outside the body, is gaining significant attention. Previous research by Kakugo and colleagues had developed swarm-type molecular robots that move individually. These robots could be assembled and disassembled as a group through external manipulation. But thanks to the constructed molecular controller, the robots can self-assemble and disassemble according to a programmed sequence.
The molecular controller initiates the process by outputting a specific DNA signal equivalent to the “assemble” command. The microtubules in the same solution, modified with DNA and propelled by kinesin molecular motors, receive the DNA signal, align their movement direction, and automatically assemble into a bundled structure. Subsequently, the controller outputs a “disassemble” signal, causing the microtubule bundles to disassemble automatically. This dynamic change was achieved through precise control by the molecular circuit, which functions like a highly sophisticated signal processor. Moreover, the molecular controller coexists with molecular robots, eliminating the need for external manipulation.
Advancing this technology is expected to contribute to the development of more complex and advanced autonomous molecular systems. As a result, molecular robots might perform tasks that cannot be accomplished alone by assembling according to commands and then dispersing to explore targets. Additionally, this research expanded the activity conditions of molecular robots by integrating different molecular groups, such as the DNA circuit system and the motor protein operating system.
“By developing the molecular controller and combining it with increasingly sophisticated and precise DNA circuits, molecular information amplification devices, and biomolecular design technologies, we expect swarm molecular robots to process a more diverse range of biomolecular information automatically,” adds Nomura. ” This advancement may lead to the realization of innovative technologies in nanotechnology and the medical field, such as nanomachines for in-situ molecular recognition and diagnosis or smart drug delivery systems.” More