Aurelia Butler

Using artificial intelligence (AI), breast radiologists in Denmark have improved breast cancer screening performance and reduced the rate of false-positive findings. Results of the study were published today in Radiology, a journal of the Radiological Society of North America (RSNA).
Mammography successfully reduces breast cancer mortality, but also carries the risk of false-positive findings. In recent years, researchers have studied the use of AI systems in screening.
“We believe AI has the potential to improve screening performance,” said Andreas D. Lauritzen, Ph.D., a post-doctoral student at the University of Copenhagen and researcher at Gentofte Hospital in Denmark.
When used to triage likely normal screening results or assist with decision support, AI also can substantially reduce radiologist workload.
“Population-based screening with mammography reduces breast cancer mortality, but it places a substantial workload on radiologists who must read a large number of mammograms, the majority of which don’t warrant a recall of the patient,” Dr. Lauritzen said. “The reading workload is further compounded when screening programs employ double reading to improve cancer detection and decrease false-positive recalls.”
Dr. Lauritzen and colleagues set out to compare workload and screening performance in two cohorts of women who underwent screening before and after AI implementation.
The retrospective study compared two groups of women between the ages of 50 and 69 who underwent biennial mammography screening in the Capital Region of Denmark.

In the first group, two radiologists read the mammograms of women screened between October 2020 and November 2021 before the implementation of AI. The screening mammograms of the second group of women performed between November 2021 and October 2022 were initially analyzed by AI. Mammograms deemed likely to be normal by AI were then read by one of 19 specialized full-time breast radiologists (called a single-read). The remaining mammograms were read by two radiologists (called a double-read) with AI-assisted decision support.
The commercially available AI system used for screening was trained by deep learning models to highlight and rate suspicious lesions and calcifications within mammograms. All women who underwent mammographic screening were followed for at least 180 days. Invasive cancers and ductal carcinoma in situ (DCIS) detected through screening were confirmed through needle biopsy or surgical specimens.
In total, 60,751 women were screened without AI, and 58,246 women were screened with the AI system. In the AI implementation group, 66.9% (38,977) of the screenings were single-read, and 33.1% (19,269) were double-read with AI assistance.
Compared to screening without AI, screening with the AI system detected significantly more breast cancers (0.82% versus 0.70%) and had a lower false-positive rate (1.63% versus 2.39%).
“In the AI-screened group, the recall rate decreased by 20.5 percent, and the radiologists’ reading workload was lowered by 33.4 percent,” Dr. Lauritzen said.
The positive predictive value of AI screening was also greater than that of screening without AI (33.5% versus 22.5%). In the AI group, a higher proportion of invasive cancers detected were 1 centimeter or less in size (44.93% vs. 36.60%).

“All screening performance indicators improved except for the node-negative rate which showed no evidence of change,” Dr. Lauritzen said.
Dr. Lauritzen said more research is needed to evaluate long-term outcomes and ensure overdiagnosis does not increase.
“Radiologists typically have access to the women’s previous screening mammograms, but the AI system does not,” he said. “That’s something we’d like to work on in the future.”
It is also important to note that not all countries follow the same breast cancer screening protocols and intervals. U.S. breast cancer screening protocols differ from protocols used in Denmark. More

Let’s say you want to train a robot so it understands how to use tools and can then quickly learn to make repairs around your house with a hammer, wrench, and screwdriver. To do that, you would need an enormous amount of data demonstrating tool use.
Existing robotic datasets vary widely in modality — some include color images while others are composed of tactile imprints, for instance. Data could also be collected in different domains, like simulation or human demos. And each dataset may capture a unique task and environment.
It is difficult to efficiently incorporate data from so many sources in one machine-learning model, so many methods use just one type of data to train a robot. But robots trained this way, with a relatively small amount of task-specific data, are often unable to perform new tasks in unfamiliar environments.
In an effort to train better multipurpose robots, MIT researchers developed a technique to combine multiple sources of data across domains, modalities, and tasks using a type of generative AI known as diffusion models.
They train a separate diffusion model to learn a strategy, or policy, for completing one task using one specific dataset. Then they combine the policies learned by the diffusion models into a general policy that enables a robot to perform multiple tasks in various settings.
In simulations and real-world experiments, this training approach enabled a robot to perform multiple tool-use tasks and adapt to new tasks it did not see during training. The method, known as Policy Composition (PoCo), led to a 20 percent improvement in task performance when compared to baseline techniques.
“Addressing heterogeneity in robotic datasets is like a chicken-egg problem. If we want to use a lot of data to train general robot policies, then we first need deployable robots to get all this data. I think that leveraging all the heterogeneous data available, similar to what researchers have done with ChatGPT, is an important step for the robotics field,” says Lirui Wang, an electrical engineering and computer science (EECS) graduate student and lead author of a paper on PoCo.

Wang’s coauthors include Jialiang Zhao, a mechanical engineering graduate student; Yilun Du, an EECS graduate student; Edward Adelson, the John and Dorothy Wilson Professor of Vision Science in the Department of Brain and Cognitive Sciences and a member of the Computer Science and Artificial Intelligence Laboratory (CSAIL); and senior author Russ Tedrake, the Toyota Professor of EECS, Aeronautics and Astronautics, and Mechanical Engineering, and a member of CSAIL. The research will be presented at the Robotics: Science and Systems Conference.
Combining disparate datasets
A robotic policy is a machine-learning model that takes inputs and uses them to perform an action. One way to think about a policy is as a strategy. In the case of a robotic arm, that strategy might be a trajectory, or a series of poses that move the arm so it picks up a hammer and uses it to pound a nail.
Datasets used to learn robotic policies are typically small and focused on one particular task and environment, like packing items into boxes in a warehouse.
“Every single robotic warehouse is generating terabytes of data, but it only belongs to that specific robot installation working on those packages. It is not ideal if you want to use all of these data to train a general machine,” Wang says.
The MIT researchers developed a technique that can take a series of smaller datasets, like those gathered from many robotic warehouses, learn separate policies from each one, and combine the policies in a way that enables a robot to generalize to many tasks.

They represent each policy using a type of generative AI model known as a diffusion model. Diffusion models, often used for image generation, learn to create new data samples that resemble samples in a training dataset by iteratively refining their output.
But rather than teaching a diffusion model to generate images, the researchers teach it to generate a trajectory for a robot. They do this by adding noise to the trajectories in a training dataset. The diffusion model gradually removes the noise and refines its output into a trajectory.
This technique, known as Diffusion Policy, was previously introduced by researchers at MIT, Columbia University, and the Toyota Research Institute. PoCo builds off this Diffusion Policy work.
The team trains each diffusion model with a different type of dataset, such as one with human video demonstrations and another gleaned from teleoperation of a robotic arm.
Then the researchers perform a weighted combination of the individual policies learned by all the diffusion models, iteratively refining the output so the combined policy satisfies the objectives of each individual policy.
Greater than the sum of its parts
“One of the benefits of this approach is that we can combine policies to get the best of both worlds. For instance, a policy trained on real-world data might be able to achieve more dexterity, while a policy trained on simulation might be able to achieve more generalization,” Wang says.
Because the policies are trained separately, one could mix and match diffusion policies to achieve better results for a certain task. A user could also add data in a new modality or domain by training an additional Diffusion Policy with that dataset, rather than starting the entire process from scratch.
The researchers tested PoCo in simulation and on real robotic arms that performed a variety of tools tasks, such as using a hammer to pound a nail and flipping an object with a spatula. PoCo led to a 20 percent improvement in task performance compared to baseline methods.
“The striking thing was that when we finished tuning and visualized it, we can clearly see that the composed trajectory looks much better than either one of them individually,” Wang says.
In the future, the researchers want to apply this technique to long-horizon tasks where a robot would pick up one tool, use it, then switch to another tool. They also want to incorporate larger robotics datasets to improve performance.
“We will need all three kinds of data to succeed for robotics: internet data, simulation data, and real robot data. How to combine them effectively will be the million-dollar question. PoCo is a solid step on the right track,” says Jim Fan, senior research scientist at NVIDIA and leader of the AI Agents Initiative, who was not involved with this work.
This research is funded, in part, by Amazon, the Singapore Defense Science and Technology Agency, the U.S. National Science Foundation, and the Toyota Research Institute. More

Researchers who had been using Fitbit data to help predict surgical outcomes have a new method to more accurately gauge how patients may recover from spine surgery.
Using machine learning techniques developed at the AI for Health Institute at Washington University in St. Louis, Chenyang Lu, the Fullgraf Professor in the university’s McKelvey School of Engineering, collaborated with Jacob Greenberg, MD, assistant professor of neurosurgery at the School of Medicine, to develop a way to predict recovery more accurately from lumbar spine surgery.
The results published this month in the journal Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, show that their model outperforms previous models to predict spine surgery outcomes. This is important because in lower back surgery and many other types of orthopedic operations, the outcomes vary widely depending on the patient’s structural disease but also varying physical and mental health characteristics across patients.
Surgical recovery is influenced by both preoperative physical and mental health. Some people may have catastrophizing, or excessive worry, in the face of pain that can make pain and recovery worse. Others may suffer from physiological problems that cause worse pain. If physicians can get a heads-up on the various pitfalls for each patient, that will allow for better individualized treatment plans.
“By predicting the outcomes before the surgery, we can help establish some expectations and help with early interventions and identify high risk factors,” said Ziqi Xu, a PhD student in Lu’s lab and first author on the paper.
Previous work in predicting surgery outcomes typically used patient questionnaires given once or twice in clinics that capture only one static slice of time.
“It failed to capture the long-term dynamics of physical and psychological patterns of the patients,” Xu said. Prior work training machine learning algorithms focus on just one aspect of surgery outcome “but ignore the inherent multidimensional nature of surgery recovery,” she added.

Researchers have used mobile health data from Fitbit devices to monitor and measure recovery and compare activity levels over time but this research has shown that activity data, plus longitudinal assessment data, is more accurate in predicting how the patient will do after surgery, Greenberg said.
The current work offers a “proof of principle” showing, with the multimodal machine learning, doctors can see a much more accurate “big picture” of all the interrelated factors that affect recovery. Proceeding this work, the team first laid out the statistical methods and protocol to ensure they were feeding the AI the right balanced diet of data.
Prior to the current publication, the team published an initial proof of principle in Neurosurgery showing that patient-reported and objective wearable measurements improve predictions of early recovery compared to traditional patient assessments. In addition to Greenberg and Xu, Madelynn Frumkin, a PhD psychological and brain sciences student in Thomas Rodebaugh’s laboratory in Arts & Sciences, was co-first author on that work. Wilson “Zack” Ray, MD, the Henry G. and Edith R. Schwartz Professor of neurosurgery in the School of Medicine, was co-senior author, along with Rodebaugh and Lu. Rodebaugh is now at the University of North Carolina at Chapel Hill.
In that research, they show that Fitbit data can be correlated with multiple surveys that assess a person’s social and emotional state. They collected that data via “ecological momentary assessments” (EMAs) that employ smart phones to give patients frequent prompts to assess mood, pain levels and behavior multiple times throughout day.
“We combine wearables, EMA -and clinical records to capture a broad range of information about the patients, from physical activities to subjective reports of pain and mental health, and to clinical characteristics,” Lu said.
Greenberg added that state-of-the-art statistical tools that Rodebaugh and Frumkin have helped advance, such as “Dynamic Structural Equation Modeling,” were key in analyzing the complex, longitudinal EMA data.

For the most recent study they then took all those factors and developed a new machine learning technique of “Multi-Modal Multi-Task Learning (M3TL)” to effectively combine these different types of data to predict multiple recovery outcomes.
In this approach, the AI learns to weigh the relatedness among the outcomes while capturing their differences from the multimodal data, Lu adds.
This method takes shared information on interrelated tasks of predicting different outcomes and then leverages the shared information to help the model understand how to make an accurate prediction, according to Xu.
It all comes together in the final package producing a predicted change for each patient’s post-operative pain interference and physical function score.
Greenberg says the study is ongoing as they continue to fine tune their models so they can take these more detailed assessments, predict outcomes and, most notably, “understand what types of factors can potentially be modified to improve longer term outcomes.”
This study was funded by grants from AO Spine North America, the Cervical Spine Research Society, the Scoliosis Research Society, the Foundation for Barnes-Jewish Hospital, Washington University/BJC Healthcare Big Ideas Competition, the Fullgraf Foundation, and the National Institute of Mental Health (1F31MH124291-01A). More

In disease research, it’s important to know gene expression and where in a tissue the expression is happening, but marrying the two sets of information can be challenging.
“Single-cell technologies, especially in the emerging field of spatial transcriptomics, help scientists see where in a tissue the genes are turned on or off. It combines information about gene activity with the exact locations within the disease tissues,” explains Fan Zhang, PhD, assistant professor of medicine with a secondary appointment in the Department of Biomedical Informatics at the University of Colorado School of Medicine.
“This is really valuable because it lets physicians and researchers see not just which genes are active, but also where they are active, which can give key insights into how different cells behave and interact in diseased conditions,” she continues.
Effectively combining location and genetic information has been a tough obstacle for researchers — until now.
Zhang and her lab developed a new computational machine learning method — called Spatial Transcriptomic multi-viEW, or “STew” for short — that enables the joint analysis of spatial variation and gene expression changes in a scalable way that can handle large amounts of cells.
This new technology may help researchers learn more about the spatial biology behind many different diseases and lead them to better treatment therapies.
A path toward an accurate target for effective treatment
The new technology is accurate in finding significant patterns that show where specific cell activities happen, which is important for understanding how cells work and how clinical tissues are structured in diseases. Zhang’s lab has already successfully applied STew on human tissues, including human brains, skin with inflammation, and breast cancer tumors.

For Zhang, who studies inflammatory diseases using computational AI tools and translational approaches, finding a good target for treatment is often a challenge, but STew could help change that.
“With inflamed joints, for example, the genes causing inflammation could be closer to the blood vessel through interacting with mesenchymal structures, or they could be farther away, but knowing that exact location and cell-cell communication patterns helps us better understand the underlying mechanisms,” she says.
By merging spatial biology and molecular diversity, STew gives researchers a new dimension in classifying patient heterogeneity.
“If you only use gene expression to classify patients, you don’t have the full picture,” Zhang says. “Once you add in spatial information, you have a more comprehensive understanding.”
“We expect STew to be effective in uncovering critical molecular and cellular signals in various clinical conditions, like different types of tumors and autoimmune disorders, opening new avenues for dysregulated immune pathways for therapeutic intervention for theses disease,” she continues.
A novel software-driven route to empowering collaboration
There’s another perk that comes with the development of STew: collaboration. Scientific discoveries often benefit from experts from different fields working together.

Because STew has a wide application, Zhang says the software will bring researchers together in new and exciting ways that will ultimately benefit the field of medicine and offer promise to patients in need of treatments.
“We want to encourage researchers across specialties, skillsets, and even departments, to collaborate in ways that they previously might not have been able to do,” Zhang says. “We can accomplish more together, so it’s important to boost data-driven and AI tool-motivated collaboration in a way that is meaningful.” More

In times of crisis, effective humanitarian aid depends largely on the fast and efficient allocation of resources and personnel. Accurate data about the locations and movements of affected people in these situations is essential for this. Researchers from the University of Tokyo, working with the World Bank, have produced a framework to analyze and visualize population mobility data, which could help in such cases.
Wars, famines, outbreaks, natural disasters … There are, sadly, many reasons why populations might be forced or feel compelled to leave their homes in search of refuge elsewhere, and these cases continue to grow. The United Nations estimated in 2023 that there were over 100 million forcibly displaced people in the world. Over 62 million of these individuals are considered internally displaced people (IDPs), those in particularly vulnerable situations due to being stuck within the borders of their countries, from which they might be trying to flee.
The circumstances that displace populations are inevitably chaotic and certainly, but not exclusively, in cases of conflict, information infrastructure can be impeded. So, authorities and agencies trying to get a handle on crises are often operating with limited data on the people they are trying to help. But the lack of data alone is not the only problem; being able to easily interpret data, so that nonexperts can make effective decisions based on it, is also an issue, especially in rapidly evolving situations where the stakes, and tensions, are high.
“It’s practically impossible to provide aid agencies and others with accurate real time data on affected populations. The available data will often be too fragmented to be useful directly,” said Associate Professor Yuya Shibuya from the Interfaculty Initiative in Information Studies. “There have been many efforts to use GPS data for such things, and in normal situations, it has been shown to be useful to model population behavior. But in times of crisis, patterns of predictability break down and the quality of data decreases. As data scientists, we explore ways to mitigate these problems and have developed a tracking framework for monitoring population movements by studying IDPs displaced in Russia’s invasion of Ukraine in 2022.”
Even though Ukraine has good enough network coverage throughout to acquire GPS data, the data generated is not representative of the entire population. There are also privacy concerns, and likely other significant gaps in data due to the nature of conflict itself. As such, it’s no trivial task to model the way populations move. Shibuya and her team had access to a limited dataset which covered the period a few weeks before and a few weeks after the initial invasion on Feb. 24, 2022. This data contained over 9 million location records from over 100,000 anonymous IDPs who opted in to share their location data.
“From these records, we could estimate people’s home locations at the regional level based on regular patterns in advance of the invasion. To make sure this limited data could be used to represent the entire population, we compared our estimates to survey data from the International Organization for Migration of the U.N.,” said Shibuya. “From there, we looked at when and where people moved just prior to and for some time after the invasion began. The majority of IDPs were from the capital, Kyiv, and some people left as early as five weeks before Feb. 24, perhaps in anticipation, though it was two weeks after that day that four times as many people left. However, a week later still, there was evidence some people started to return.”
That some people return to afflicted areas is just one factor that confounds population mobility models — in actual fact, people may move between locations, sometimes multiple times. Trying to represent this with a simple map with arrows to show populations could get cluttered fast. Shibuya’s team used color-coded charts to visualize its data, which allow you to see population movements in and out of regions at different times, or dynamic data, in a single image.
“I want visualizations like these to help humanitarian agencies gauge how to allocate human resources and physical resources like food and medicine. As they tell you about dynamic changes in populations, not just A to B movements, I think it could mean aid gets to where it’s needed and when it’s needed more efficiently, reducing waste and overheads,” said Shibuya. “Another thing we found that could be useful is that people’s migration patterns vary, and socioeconomic status seems to be a factor in this. People from more affluent areas tended to move farther from their homes than others. There is demographic diversity and good simulations ought to reflect this diversity and not make too many assumptions.”
The team worked with the World Bank on this study, as the international organization could provide the data necessary for the analyses. They hope to look into other kinds of situations too, such as natural disasters, political conflicts, environmental issues and more. Ultimately, by performing research like this, Shibuya hopes to produce better general models of human behavior in crisis situations in order to alleviate some of the impacts those situations can create. More

A research group at Osaka University has developed an innovative positioning system, correctly aligning the coordinates of the real and virtual worlds without the need to define routes in advance. This is achieved by integrating two vision-based self-location estimation methods: visual positioning systems (VPS) and natural feature-based tracking. This development will lead to the realization of versatile drone-based mixed reality (MR) using drones available on the market. Drone-based MR is expected to see use in a variety of applications in the future, such as urban landscape simulation and support for maintenance and inspection work, contributing to further development of drone applications, especially in the fields of architecture, engineering, and construction (AEC).
In recent years, there has been a growing interest in the integration of drones across diverse sectors, particularly within AEC. The use of drones in AEC has expanded due to their superior features in terms of time, accuracy, safety, and cost. The amalgamation of drones with MR stands out as a promising avenue as it is not restricted by the user’s range of action and is effective when performing landscape simulations for large-scale spaces such as cities and buildings. Previous studies proposed methods to integrate MR and commercial drones using versatile technologies such as screen sharing and streaming delivery; however, these methods required predefined drone flight routes to match the movements of the real and virtual world, thus reducing the versatility of the application and limiting use cases of MR.
While this research does not implement a drone-based MR application for actual use, the proposed alignment system is highly versatile and has the potential for various additional functionalities in the future. This brings us one step closer to realizing drone-centric MR applications that can be utilized throughout the entire lifecycle of architectural projects, from the initial stages of design and planning to later stages such as maintenance and inspection.
First author Airi Kinoshita mentions, “The integration of drones and MR has the potential to solve various social issues, such as those in urban planning and infrastructure development and maintenance, disaster response and humanitarian aid, cultural protection and tourism, and environmental conservation by freeing MR users from the constraints of experiencing only their immediate vicinity, enabling MR expression from a freer perspective.” More

Biological processes depend on puzzle pieces coming together and interacting. Under specific conditions, these interactions can create something new without external input. This is called self-organization, as seen in a school of fish or a flock of birds. Interestingly, the mammalian embryo develops similarly. In PNAS, David Brückner and Gašper Tkačik from the Institute of Science and Technology Austria (ISTA) introduce a mathematical framework that analyzes self-organization from a single cell to a multicellular organism.
When an embryo develops, many types of cells with different functions need to be generated. For example, some cells will become part of the eye and record visual stimuli, while others will be part of the gut and help digest food. To determine their roles, cells are constantly communicating with each other using chemical signals.
Thanks to this communication, during development, everything is well synchronized and coordinated, and yet there is no central control responsible for this. The cell collective is self-organized and orchestrated by the interactions between the individuals. Each cell reacts to signals of its neighbors. Based on such self-organization, the mammalian embryo develops from a single fertilized egg cell into a multicellular organism.
David Brückner and Gašper Tkačik from the Institute of Science and Technology Austria (ISTA) have now established a mathematical framework that helps analyze this process and predict its optimal parameters. Published in PNAS, this approach represents a unifying mathematical language to describe biological self-organization in embryonic development and beyond.
The self-assembling embryo
In nature, self-organization is all around us: we can observe it in fish schools, bird flocks, or insect collectives, and even in microscopic processes regulated by cells. NOMIS fellow and ISTA postdoc David Brückner is interested in getting a better understanding of these processes from a theoretical standpoint. His focus lies on embryonic development — a complex process governed by genetics and cells communicating with each other.
During embryonic development, a single fertilized cell turns into a multicellular embryo containing organs with lots of different features. “For many steps in this developmental process, the system has no extrinsic signal that directs it what to do. There is an intrinsic property of the system that allows it to establish patterns and structures,” says Brückner. “The intrinsic property is what is known as self-organization.” Even with unpredictable factors — which physicists call “noise” — the embryonic patterns are formed reliably and consistently. In recent years, scientists have gained a deeper understanding of the molecular details that drive this complex process. A mathematical framework to analyze and quantify its performance, however, was lacking. The language of information theory provides answers.

Bridging expertise
“Information theory is a universal language to quantify structure and regularity in statistical ensembles, which are a collection of replicates of the same process. Embryonic development can be seen as such a process that reproducibly generates functional organisms that are very similar but not identical,” says Gašper Tkačik, professor at ISTA and expert in this field. For a long time, Tkačik has been studying how information gets processed in biological systems, for instance in the fly embryo. “In the early fly embryo, patterns are not self-organized,” he continues. “The mother fly puts chemicals into the egg that instruct the cells on what actions to take.” As the Tkačik group had already developed a framework for this system, Brückner reached out to develop one for the mammalian embryo as well. “With Gašper’s expertise in information theory, we were able to put it together,” Brückner adds excitedly.
Beyond embryo development?
During embryonic development, cells exchange signals and are constantly subject to random, unpredictable fluctuations (noise). Therefore, cellular interactions must be robust. The new framework measures how these interactions are possibly optimized to withstand noise. Using computer simulations of interacting cells, the scientists explored the conditions under which a system can still have a stable final result despite introducing fluctuations.
Although the framework has proven to be successful on three different developmental models that all rely on chemical and mechanical signaling, additional work will be required to apply it to experimental recordings of developmental systems. “In the future, we want to study more complex models with more parameters and dimensions,” Tkačik says. “By quantifying more complex models, we could also apply our framework to experimentally measured patterns of chemical signals in developing embryos,” adds Brückner. For this purpose, the two theoretical scientists will team up with experimentalists. More

An international research team around Marcus Sperling, a researcher at the Faculty of Physics, University of Vienna, has sparked interest in the scientific community with pioneering results in quantum physics: In their current study, the researchers reinterpret the Higgs mechanism, which gives elementary particles mass and triggers phase transitions, using the concept of “magnetic quivers.” The work has now been published in the journal “Physical Review Letters.”
The foundation of Marcus Sperling’s research, which lies at the intersection of physics and mathematics, is Quantum Field Theory (QFT) — a physical-mathematical concept within quantum physics focused on describing particles and their interactions at the subatomic level. Since 2018, he has developed the so-called “magnetic quivers” along with colleagues — a graphical tool that summarizes all the information needed to define a QFT, thus displaying complex interactions between particle fields or other physical quantities clearly and intuitively.
Metaphorical Magnetic Quivers
A quiver consists of directed arrows and nodes. The arrows represent the quantum fields (matter fields), while the nodes represent the interactions — e.g., strong, weak, or electromagnetic — between the fields. The direction of the arrows indicates how the fields are charged under the interactions, e.g., what electric charge the particles carry. Marcus Sperling explains, “The term ‘magnetic’ is also used metaphorically here to point to the unexpected quantum properties that are made visible by these representations. Similar to the spin of an electron, which can be detected through a magnetic field, magnetic quivers reveal certain properties or structures in the QFTs that may not be obvious at first glance.” Thus, they offer a practical way to visualize and analyze complex quantum phenomena, facilitating new insights into the underlying mechanisms of the quantum world.
Supersymmetric QFTs
For the current study, the stable ground states (vacua) — the lowest energy configuration in which no particles or excitations are present — in a variety of “supersymmetric QFTs” were explored. These QFTs, with their simplified space-time symmetry, serve as a laboratory environment, as they resemble real physical systems of subatomic particles but have certain mathematical properties that facilitate calculations. FWF START award winner Sperling said, “Our research deals with the fundamentals of our understanding of physics. Only after we have understood the QFTs in our laboratory environment can we apply these insights to more realistic QFT models.” The concept of magnetic quivers — one of the main research topics of Sperling’s START project at the University of Vienna — was used as a tool to provide a precise geometric description of the new quantum vacua.
Decay & Fission: Higgs Mechanism Reinterpreted
With calculations based on linear algebra, the research team demonstrated that — analogous to radioactivity in atomic nuclei — a magnetic quiver can decay into a more stable state or fission into two separate quivers. These transformations offer a new understanding of the Higgs mechanism in QFTs, which either decay into simpler QFTs or fission into separate, independent QFTs. Physicist Sperling stated, “The Higgs mechanism explains how elementary particles acquire their mass by interacting with the Higgs field, which permeates the entire universe. Particles interact with this field as they move through space — similar to a swimmer moving through water.” A particle that has no mass usually moves at the speed of light. However, when it interacts with the Higgs field, it “sticks” to this field and becomes sluggish, leading to the manifestation of its mass. The Higgs mechanism is thus a crucial concept for understanding the fundamental building blocks and forces of the universe. Mathematically, the “decay and fission” algorithm is based on the principles of linear algebra and a clear definition of stability. It operates autonomously and requires no external inputs. The results achieved through physics-inspired methods are not only relevant in physics but also in mathematical research: They offer a fundamental and universally valid description of the complex, intertwined structures of the quantum vacua, representing a significant advance in mathematics. More