Thursday, July 02, 2026

 

Revolutionizing software quality: new study explores large language models' pioneering role in defect detection




Higher Education Press
image 

image: 

Classification of software defect detection studies using LLM

view more 

Credit: HIGHER EDUCATION PRESS

 





The ongoing evolution of software defect detection methodologies leveraging large language models is rapid; however, the current research landscape has not been adequately investigated. Existing reviews inclusively categorize smaller language models as LLMs, failing to concentrate specifically on the domain of software defect detection and omitting recent advancements in the field. This highlights a gap in the detailed analysis and assessment of cutting-edge applications of large language models in detecting software flaws.

To solve the problems, a research team led by Zulie PAN published their new research on 15 June 2026 in Frontiers of Computer Science co-published by Higher Education Press and Springer Nature.

The team collected high-quality research papers focusing on LLM-based software defect detection, study the progress of related technology development, and predict the development prospects.

In the research, they categorize and summarize existing research based on the distinct applications of LLMs in dynamic and static detection scenarios. Firstly, dynamic detection methods are categorized based on the different phases in which they employ LLMs, such as using them for test case generation, providing feedback guidance, and conducting output assessment. Secondly, static detection methods are classified according to whether they analyze the source code or the binary of the software under test.

They investigated the prompt engineering and model fine-tuning strategies adopted within collected studies. Furthermore, they analyzed the LLMs utilized in the collected papers, the datasets employed, the target software addressed, and the technical performance outcomes. Finally, they summarize the overall concepts and trends of existing research and propose potential directions for future LLM-based software defect detection studies. As the technology evolves, LLMs promise to not only streamline the defect detection process but also to usher in a new era of efficient and error-free software development, thereby elevating industry standards and safeguarding user experiences. This study marks a significant milestone in leveraging the power of language intelligence to mitigate software risks, setting the stage for future innovations in the realm of software engineering.

 

Ants: Who looks after the injured in a colony?




University of Würzburg

Ants during an amputation 

image: 

Carpenter ants perform amputations as a preventive measure. Their motto is: Better safe than sorry when it comes to potential infections.

view more 

Credit: Bart Zijlstra





Patients in hospital generally trust the nursing staff. After all, they have undergone training and, in some cases, have several years of professional experience.

In the case of carpenter ants, it is not nursing expertise that determines who cares for the patients: “There are no specialised ‘medics’ in the colonies. Instead, this task is carried out by worker ants that are in the process of transitioning from brood care to foraging,” says Dr Erik Frank, senior author of the study and head of an Emmy Noether research group at the Chair of Animal Ecology and Tropical Biology at the University of Würzburg. This transitional phase for the ants usually lasts 20 days.

Another crucial factor is how many previous interactions the carer has had with the injured ant, says Alba Motes-Rodrigo, co-author from the University of Lausanne (Switzerland). These include social interactions such as grooming one another or encountering each other by chance in the nest and touching each other with their antennae.

Ants transitioning between indoor and outdoor duties roam throughout the entire nest and are therefore better connected than other members of their colony. The research team published their findings in the journal *Proceedings of the National Academy of Sciences* (PNAS).

Fully automated tracking: 660 ants under scrutiny

For the study, the team examined six colonies, each comprising 110 ants of the species Camponotus fellah, which belongs to the genus of carpenter ants and is found primarily in the Middle East. Using a fully automated tracking system, the researchers were able to precisely monitor the movements and hundreds of thousands of interactions of each ant, as well as their wound care, over a period of weeks.

“We have long known that the spatial organisation of a colony governs everyday tasks such as brood care or foraging. But our findings go even further,” explains Dr Ebi George, a co-author from the University of Lausanne. They show that the everyday spatial and social overlap between workers also determines temporary tasks such as life-saving wound care, George elaborates.

Carpenter ants: Masters of amputation

In a previous study, Frank and his team had already observed how carpenter ants treat wounds: they amputate the injured legs of their fellow ants by biting them off and treating them with antimicrobial substances. Their guiding principle is always: better safe than sorry.

The ants carry out prophylactic amputations. This not only protects the colony from infection but also doubles the survival rate of the injured workers. “Back then, we were able to show how the wounds are treated. Our current study now reveals who is primarily responsible for this,” says Frank.

 

Bath researchers join NASA Earth System Explorers mission





University of Bath






Bath researchers Dr Neil Hindley and Professor Corwin Wright, from the University of Bath’s Centre for Climate Adaptation & Environment Research (CAER), have joined the science team behind NASA’s next-generation satellite mission.  

The STRIVE mission, which stands for Stratosphere Troposphere Response using Infrared Vertically-resolved light Explorer, will take high-resolution, daily measurements of global temperature, aerosol pollution and ozone levels throughout the Earth’s atmosphere. These observations will help improve longer-range weather and air quality forecasts, and strengthen early warnings for extreme weather, particularly for vulnerable coastal communities, where almost half of the world’s population lives. 

The Bath team will focus on understanding atmospheric waves, the ripples of energy that travel around Earth and drive our daily weather and global climate. By understanding and predicting how these waves transport energy and momentum, researchers hope to better predict environmental impacts such as extreme weather and long-term climate projections.  

Dr Hindley, a Senior Research Fellow in CAER, explained: “The atmosphere is made up of several, deeply interconnected layers. As a rule of thumb, the further forward in time you want to predict, the more of the atmosphere you need to understand and simulate – and this usually means going upwards. 

“STRIVE will be able to see these waves in full 3-D for the first time. Our role in the Science Team is to analyse these wave signatures and measure their momentum transport, breaking and deposition throughout the middle atmosphere, using cutting-edge 3-D spectral analysis techniques pioneered at Bath.” 

STRIVE measurements will provide unprecedented insights into the dynamics and chemistry of the middle atmosphere, the dynamic region 10 to 100 km above the Earth’s surface. Understanding how it circulates will help forecasters create better numerical climate simulations, improving how far ahead we can reliably predict weather and climate, potentially extending forecasting from weeks to full seasons. 

As part of the development phase of the mission, the Bath team tested how STRIVE will observe the atmosphere once in orbit. 

Dr Hindley added: “In collaboration with our US partners, we flew a virtual satellite through an ultra-high-resolution model of the atmosphere, sampling fields such as temperature exactly as STRIVE will see it. This gave us a suite of synthetic measurements to develop our 3D analysis tools, so that we’re ready for the real data after launch.” 

The observations will also help improve modelling of the upper atmosphere, where satellites orbit the Earth. Here, at the edge of space, volatile atmospheric conditions can cause sudden and dramatic changes in the drag experienced by satellites, which can have serious consequences, such as whole constellations of satellites de-orbiting shortly after launch. Improved understanding of the causes of these conditions from STRIVE observations will lead to improved forecast modelling, which will support safer and more reliable satellite operations.  

The mission is led by Professor Lyatt Jaeglé at the University of Washington in Seattle, and is one of two new Earth System Explorers missions selected by NASA to improve understanding of the atmosphere and support faster, more informed decision-making during extreme weather events.  

Professor Jaeglé said: “The STRIVE mission will help us close one of the biggest blind spots in our observing system, the middle atmosphere. By resolving it in unprecedented detail, we’ll improve our ability to predict extreme weather, track pollution and wildfire smoke, and monitor the recovery of the ozone layer. 

“We greatly value the scientific contributions from our UK partners at the University of Bath, Neil Hindley and Corwin Wright. They bring key deep expertise in gravity wave science and are pioneers in spectral 3D analysis of the type of data that STRIVE will provide.” 

Both missions are currently in development and will undergo further review in 2027. If approved, they are expected to launch from 2030.  

ENDS 

Notes to editors:   

For more information, please contact Sarah Baker-Gaunt at the University of Bath Press Office on press@bath.ac.uk 

About the University of Bath 

The University of Bath is one of the UK's leading universities, recognised for high-impact research, excellence in education, an outstanding student experience and strong graduate prospects. 

  • We are ranked among the top 10% of universities globally, placing 125th in the QS World University Rankings 2027. 

  • We are ranked in the top 10 in all of the UK’s major university guides. 

  • The University achieved a triple Gold award in the last Teaching Excellence Framework 2023, the highest awards possible, for both the overall assessment and for student outcomes and student experience. The Teaching Excellence Framework (TEF) is a national scheme run by the Office for Students (OfS). 

  • We are The Times and The Sunday Times Sport University of the Year 2026. 

Research at Bath is shaping a better future through innovation in sustainability, health, and digital technologies. Find out all about our Research with Impact: http://bit.ly/3ISz1Wu  

 

Modern life may be outpacing the human mind




Singapore University of Technology and Design






The human brain evolved for a world of familiar faces, immediate threats and small social groups. But the world around us is changing far faster than human biology can keep pace. That mismatch may help explain some of the stress, loneliness and constant comparison people experience today.

The review, co-authored by Dr Jose Yong, Senior Lecturer at James Cook University, Singapore, and Dr Sarah Chan, Research Fellow at the Lee Kuan Yew Centre for Innovative Cities at SUTD, is published in Behavioral Sciences. Titled Evolutionary mismatch, stress, and competition: Making sense of psychosocial problems in the polycrisis era, it examines how stress, competition and loneliness can be understood through an evolutionary lens.

Evolutionary mismatch describes what happens when human instincts shaped in one kind of environment are forced to operate in a very different one. Humans evolved in smaller, close-knit groups, where danger, belonging, status and trust were read through familiar people and everyday face-to-face signals. Now, those same instincts are being triggered in dense cities, digital platforms, unequal societies and a world shaped by overlapping pressures. The result is an internal confusion: responses that once made sense in a small familiar group can feel out of place, or simply overwhelming, in modern life.

Social media makes this mismatch especially visible. The urge to understand our place within a group may once have helped people maintain trust and cooperation among familiar faces. Today, that same instinct can be triggered by an endless stream of curated lives, achievements and status signals.

At the centre of the paper is competition. Modern environments can intensify the feeling that others are judging, outperforming or leaving us behind. The authors propose that this heightened sense of competition may be one pathway through which evolutionary mismatch contributes to stress and poorer wellbeing.

“Competition is not new, but modern life can make it feel constant,” said Dr Yong. “An evolutionary perspective may help explain why people respond so strongly to comparison and the fear of falling behind, even when those signals come from strangers or screens rather than a small social group.”

The paper draws on existing research and theory rather than new data. It presents evolutionary mismatch as one way of understanding modern social and psychological problems, alongside psychological, social and economic explanations. These ideas will need to be tested through real-world research.

That matters because the response to modern stress cannot rest only on telling individuals to be more resilient. If environments are activating old instincts in new and unhelpful ways, then cities, workplaces, digital platforms and communities also need to be part of the solution.

For SUTD, the work connects closely with human-centred design and urban wellbeing. Density alone does not determine how people feel in a city. What matters is whether a place feels crowded, threatening or difficult to navigate. Greener surroundings, stronger community ties and more thoughtful social design may help ease those pressures without requiring cities to become less dense.

“Stress, loneliness and anxiety are often treated as personal or lifestyle problems,” said Dr Chan. “But they may also reflect a mismatch between the environments people live in and the conditions our minds and bodies evolved to navigate. That means we should think not only about individual resilience, but also about how cities and communities are designed.”

Future studies could examine how perceived competition and wellbeing vary across greener neighbourhoods, places that feel more or less crowded, communities with different levels of social connection, and digital spaces that encourage or reduce comparison.

None of this is an argument for returning to a simpler past, or a suggestion that modern life is inherently broken. It is a case for designing the present more thoughtfully. Understanding where modern life conflicts with the conditions human beings evolved to navigate could help researchers, designers and policymakers create cities and communities that feel less alienating and more supportive of everyday wellbeing.

“We need to design interventions that work with rather than against our evolved human nature,” said Dr Yong.

 

 

Most Europeans see economic growth as essential for a sustainable future




Universitat Autonoma de Barcelona






Most Europeans citizens consider economic growth a necessary condition, rather than an obstacle, for advancing towards a sustainable future. This is the conclusion reached in a new study by the Institute of Environmental Science and Technology of the Universitat Autònoma de Barcelona (ICTA-UAB), published in Nature Communications.

In a context shaped by debates on climate change and sustainability, a study based on a survey of 17,000 citizens across 13 European countries analyses how citizens perceive the role of economic growth in building a sustainable society. Sixty percent of participants consider it essential for achieving key social and environmental goals, including environmental protection, public services, economic stability, and life satisfaction. By contrast, fewer than 10% expressed views that were clearly sceptical about economic growth.

“Discussions about post-growth and degrowth have become increasingly prominent in academic and policy circles,” says Professor Ivan Savin, lead author of the study. “Our findings show that these debates are not yet reflected in public opinion. Most citizens still view economic growth as an important means of achieving broader societal goals.”

The research challenges the common assumption that support for growth comes at the expense of environmental concern. People with pro-growth views are no less concerned about climate change than those who are sceptical about growth. Overall, they are also no less supportive of climate action. Instead, many respondents view economic growth and sustainability as compatible objectives rather than competing priorities.

According to the findings, support for growth is associated not only with values linked to personal success and material wellbeing, but also with equality and collective welfare. According to the researchers, this suggests that many Europeans see economic growth as a tool for improving society as a whole rather than a means of increasing personal prosperity.

Views also varied across countries. Citizens in wealthier and more equal societies tend to be less supportive of economic growth, suggesting that they feel less dependent on further economic expansion when basic needs and living standards are already widely secured.

The findings have important implications for climate policy. While growth-sceptical perspectives are gaining visibility in academic discussions, public support remains centred on the belief that economic growth can help deliver environmental and social progress. Policymakers seeking to accelerate sustainability transitions should therefore connect climate policies more clearly with citizens’ expectations regarding jobs, economic security and quality of life.

 

From infancy to old age, nationwide China study maps how serious lung infections change across life stages



Analysis of 695,142 hospitalized patients identifies eight life-stage pathogen profiles, three co-detection networks, and marked male predominance in bacterial and fungal infections



Science China Press

How lower respiratory infection risks shift across life stages in hospitalized patients in China 

image: 

This graphic summarizes a nationwide analysis of 695,142 bronchoalveolar lavage fluid samples from hospitalized patients with lower respiratory infections in China. It shows eight life-stage pathogen profiles, three characteristic co-detection networks built from 124 significant pathogen pairs, and sex-based differences in hospitalization patterns, including higher male representation for several bacterial and fungal pathogens.

view more 

Credit: ©Science Bulletin





BEIJING - Serious lower respiratory infections do not follow the same pathogen rules at every age. A nationwide study in China has now shown that the microbes most often detected in hospitalized patients shift dramatically from infancy to old age, and that these shifts are accompanied by distinct co-detection networks and clear sex disparities.

The study, published in Science Bulletin, analyzed 695,142 bronchoalveolar lavage fluid samples from hospitalized patients with lower respiratory infections collected between January 2022 and May 2025. The samples came from 4,758 hospitals in 306 cities across all 31 provincial-level regions of mainland China. Using targeted next-generation sequencing, the researchers tested all samples for 28 common respiratory pathogens under a unified workflow, making this one of the largest standardized pathogen-mapping studies of hospitalized lower respiratory infections to date.

At least one of the 28 target pathogens was detected in 82.09% of samples. Overall, the most prevalent pathogens were Mycoplasma pneumoniae, Haemophilus influenzae, and Klebsiella pneumoniae. But the biggest message from the study was that different life stages faced very different pathogen landscapes. Instead of relying on broad conventional age brackets, the team used real-world pathogen data to define eight consecutive life-stage groups: infant, toddler, child, adolescent, young adult, middle-aged, elderly, and octogenarian.

The contrasts were striking. In infants, Pneumocystis jirovecii, rhinovirus, and respiratory syncytial virus together accounted for 62.59% of hospitalizations. In toddlers, children, and adolescents, Mycoplasma pneumoniae became the dominant detected pathogen, reaching 66.02% in children. As age increased, the spectrum shifted toward bacteria, with Klebsiella pneumoniae, Acinetobacter baumannii, and Pseudomonas aeruginosa becoming more prominent in the oldest patients. In other words, the main pathogens linked to hospitalization were not the same across the lifespan.

The study also found that pathogens did not simply appear one by one. Among 378 possible pathogen pairs, 124 showed significant positive co-detection relationships. These pairs clustered into three characteristic networks: a pediatric-associated network, a pan-age network dominated by fungi and certain viral co-detections, and an elderly-associated network. Nearly 80% of observed co-detections occurred within, rather than between, these networks. This pattern suggests that when one pathogen is detected, clinicians may need to pay closer attention to a predictable set of co-detected partners rather than treating co-detection as random.

Sex differences were another major finding. Male hospitalized cases outnumbered female cases by 1.58 to 1, corresponding to 156,288 excess male hospitalizations in the study population. This imbalance was driven mainly by bacterial and fungal pathogens rather than by viruses. Bacteria accounted for 48.14% of the excess male hospitalizations, and fungi accounted for 17.89%. Pathogens such as Klebsiella pneumoniae, Pneumocystis jirovecii, and Acinetobacter baumannii were detected more frequently in male patients. Age mattered here as well: males were overrepresented among hospitalized patients at ages 0 to 4 years, the gap narrowed between ages 5 and 35 years, and then widened again after age 35.

Together, these findings provide a more detailed picture of how hospitalized lower respiratory infections vary across life stages and sexes. The researchers say the results may help public health teams refine age- and sex-stratified prevention strategies, support developers in designing life-stage-specific diagnostic panels, and assist clinicians in anticipating co-detection patterns in hospitalized patients.

To support real-world use, the team also developed an interactive online platform that allows clinicians and public health practitioners to query age- and sex-stratified pathogen detection data and co-detection patterns from hospitalized lower respiratory infection cases nationwide: https://psi.tracepatho.com/

The authors note that the study reflects hospitalized patients who underwent bronchoalveolar lavage fluid sampling and targeted sequencing in routine care. As a result, the dataset may overrepresent more severe or diagnostically complex cases and should not be taken as a direct proxy for all lower respiratory infections seen in outpatient settings.

The study was led by Yamin Sun, together with Min Wang, Tian Qin, Jue Liu, Jiale Yuan, Pei Li, Xinchang Lun, Zhigang Wang, Linghang Wang, Rui Song, Xi Wang, Ronghua Jin, and Jianguo Xu. Ronghua Jin and Jianguo Xu are the co-corresponding authors. The work was supported by the National Key Research and Development Program of China and the Tengfei Initiative of the National Key Laboratory of Intelligent Tracking and Forecasting for Infectious Diseases.

About Beijing Ditan Hospital, Capital Medical University:

Beijing Ditan Hospital, Capital Medical University, established in 1946, is a top-tier comprehensive hospital specializing in infectious diseases. Beijing Ditan Hospital is the National Center for Infectious Diseases (Beijing) and is also a China Regional Medical Center. With major clinical, teaching, and research responsibilities in infectious diseases and related disciplines, the hospital plays a leading role in the diagnosis, treatment, and translational research of infectious diseases in China.

About the National Key Laboratory of Intelligent Tracking and Forecasting for Infectious Diseases:

The National Key Laboratory of Intelligent Tracking and Forecasting for Infectious Diseases is a national research platform relying on China CDC and co-established with Nankai University and Beijing Ditan Hospital, Capital Medical University. The laboratory integrates expertise in pathogen tracing, outbreak surveillance, early warning, data science, and intelligent decision-making to support the prevention and control of infectious diseases. Its work focuses on building advanced technologies and interdisciplinary research capacity for infectious disease monitoring, prediction, and public health response.

 

A subtle difference offers insight into bacteria survival strategies



New research shows not all bacteria behave the same as E. coli


University of California - San Diego





Escherichia coli (E. coli) are mostly harmless bacteria that live in the intestines of animals and humans. They are the most well-studied bacteria and, often, when scientists discover something about E. coli, they extrapolate that discovery across all bacteria. So when scientists learned that E. coli allocates its resources to grow as fast as the environment allows, it was assumed all bacteria behaved similarly.

But researchers at the University of California San Diego have discovered that Bacillus subtilis (B. subtilis), a bacterium commonly found in soil, employs a different survival strategy. The result, published in Science, raises the question of whether other types of bacteria use alternative strategies, and how that knowledge might help researchers think differently about antibiotic tolerance (see Sidebar for more information).

Slower Growth for Better Survival

E. coli grow as fast as conditions allow — sometimes doubling every 20 minutes — so even in adverse conditions, the bacteria continue to multiply as quickly as possible. From an evolutionary point of view, it makes sense: the more E. coli there are, the better the odds that some will survive.

A decade ago, researchers in UC San Diego Professor of Physics Suckjoon Jun's lab set out to reproduce a landmark E. coli result in B. subtilis, and got a surprise. In E. coli, when protein production is partially blocked with an antibiotic, the cell compensates by building even more ribosomes, its protein-making machinery. The team expected B. subtilis to respond the same way. Instead, its ribosome levels stayed flat. Jun assumed they had run the experiment incorrectly, but when the result held up across repeated trials, he realized B. subtilis was managing stress in a fundamentally different way.

To learn more, Jun enlisted the help of Jade Wang, professor of bacteriology at University of Wisconsin-Madison. Wang is an expert on bacterial stringent response, particularly with E. coli and B. subtilis. Stringent response is a survival mechanism used by bacteria to adapt to harsh environmental conditions, such as nutrient deprivation or the presence of antibiotics.

Through their collaboration, Jun realized the two bacteria employ different survival control mechanisms. “Bacteria are generally thought to grow as fast as their available nutrients allow, by carefully balancing how they invest their resources. However, we found that under antibiotic stress, B. subtilis does the opposite, deliberately holding its growth below what it is capable of,” he said.

In order to thrive, bacteria must allocate resources between ribosomes, which build proteins, and the rest of the cell’s components which make the building blocks for those proteins, including amino acids. In E. coli, a small molecule called (p)ppGpp acts like a control switch: when amino acid supplies drop in unfavorable conditions, (p)ppGpp increases and tells the cell to make fewer ribosomes, keeping everything in sync.

B. subtilis has a different control switch, called guanosine triphosphate, or GTP. GTP plays two roles: it powers core processes such as protein synthesis and it acts as a regulatory signal governing functions like the stress response. During adverse environmental conditions, GTP levels in B. subtilis fall. This drop slows the production of amino acids, but the levels of ribosomes stay the same. The result is a decoupling, where the cell’s “factory” is still full of machines, but the supply of raw materials has dwindled.

Keeping the ribosome count steady when GTP drops slows bacteria growth, but also makes it more tolerant of stress. When GTP is high, growth speeds up, but the cells are more vulnerable. This trade‑off lets B. subtilis choose between faster growth and stress resistance depending on environmental conditions.

In experiments by the Jun and Wang labs, B. subtilis was much better at surviving in antibiotic environments than E. coli. This may be related to persistence, in which a small fraction of cells survives antibiotic exposure without becoming genetically resistant, then resume growing once conditions improve.

"We tend to assume bacteria are built to grow as fast as they possibly can. What surprised us is that this one chooses not to. It holds itself back to stay alive under stress, and when we flipped the switch that controls that decision, it grew faster but became far more vulnerable,” stated Jun. “The cell is constantly taking a gamble between growing and surviving, and that gamble may be part of why bacteria are so hard to kill."

The work challenges a longstanding assumption that bacteria are wired simply to grow as fast as possible, showing instead that many balance growth against survival as an active strategy. Because the slowdown is tied directly to antibiotic survival, the findings point to a new way of thinking about how bacteria tolerate antibiotics and survive stress.

Authors include Ryan Thiermann, Aniket Zodage, Taylor Rytlewski, Fangzhou Xiao, John T. Sauls, Sarah Cox, Zulfar Ghulam-Jelani, Victoria Castillo, and Suckjoon Jun  (all UC San Diego); Jin Yang, Fukang She, Danny K. Fung, Quinn A. Paulsen, David M. Stevenson, Daniel Amador-Noguez, and Jue D. Wang (all University of Wisconsin-Madison); Farshad Abdollah-Nia and James R. Williamson (both Scripps Research).

This research was funded by the National Science Foundation (1715710), the National Institutes of Health (R35-GM-136412, R35GM127088 and R35GM139622) and the Simons Foundation (SFI-PD-Pivot Fellow-00008375).