Tech & AI

Researchers found that knowledge graphs—AI systems that connect data from multiple sources—are becoming essential tools for managing traffic, parking, and city development. For urban planners and municipal leaders, this means better decisions about infrastructure investments and faster problem-solving across traditionally siloed city departments.EN

Researchers have developed mathematical models that help manufacturers optimize how robots and humans work together on assembly lines, reducing equipment and labor costs while maintaining production speed. The breakthrough addresses a real-world problem factories face as they adopt collaborative robots: figuring out the cheapest, most efficient way to divide tasks between human workers and machines.EN

Researchers have cracked a stubborn problem in federated learning—how to train AI models across multiple cell towers without exposing sensitive user data. The breakthrough improves prediction accuracy by 6% while keeping privacy intact, a win for carriers deploying AI at the network edge without massive data transfers to centralized servers.EN

Researchers have embedded electrical sensors into a lab-grown gut system to instantly measure how well the intestinal barrier is functioning. The breakthrough could accelerate development of probiotics, pharmaceuticals, and dietary products by providing faster, more precise data on how they affect gut health.EN

Researchers have designed fluorinated molecules that revive the effectiveness of meropenem, a critical antibiotic, against drug-resistant bacteria that cause difficult-to-treat infections. The compounds reduce bacterial resistance up to 64-fold, offering a potential near-term strategy to extend the clinical life of existing antibiotics while new drugs remain in development.EN

Researchers have developed a faster way to determine when solutions to complex equations will oscillate unpredictably—a capability with direct applications to engineering stability analysis, financial forecasting, and control systems. The breakthrough simplifies calculations that engineers and data scientists currently rely on, potentially cutting analysis time and improving system reliability across industries.EN

Researchers created a low-cost, customizable alternative to expensive commercial lab equipment used to test how drugs cross intestinal barriers. The 3D-printed device maintains 97% cell viability and could accelerate pharmaceutical development while reducing per-unit testing costs by orders of magnitude.EN

Researchers have developed a machine learning technique that allows medical AI systems to adapt to new imaging data without storing sensitive patient information—a critical requirement under data protection laws like GDPR. The approach could accelerate deployment of AI diagnostic tools across hospitals while keeping data risks minimal.EN

Researchers have developed a faster computational method for simulating how electrical signals spread through living tissue—critical for understanding cardiac arrhythmias and neurological disorders. The breakthrough could accelerate drug development and medical device design by making large-scale tissue simulations practical for the first time.EN

Researchers have developed a fast computational model that accurately predicts how wind turbine wakes affect downstream power generation across large farms. The breakthrough could help energy companies optimize turbine placement and operations, reducing costly trial-and-error approaches and improving renewable energy output.EN

Researchers have created the first publicly available AI classifiers that accurately predict which drugs will cause heart damage, using chemical properties and protein-target data. The breakthrough could help pharma companies avoid costly postmarket withdrawals—currently accounting for 10-14% of drug failures—and accelerate safer medicines to patients.EN

Researchers have developed a faster computational method for simulating fluid flow around obstacles—a critical task in designing everything from aircraft to offshore structures. The technique, tested on flow past cylinders, cuts computation time while maintaining accuracy, potentially accelerating product development cycles for industries reliant on complex fluid dynamics modeling.EN

Researchers have developed simpler mathematical criteria to forecast when complex systems will oscillate or destabilize—without the restrictive conditions previous methods required. The findings could streamline engineering design across telecommunications, energy grids, and industrial controls by making predictive models faster and cheaper to validate.EN

Researchers have developed a method that makes computer vision systems for counting crowds dramatically more accurate, even when trained on imperfectly labeled data. The approach could improve safety planning and resource allocation in retail, transportation, and public venues—and works with any existing AI model.EN

Researchers simulated how Swedish homes could reduce peak electricity demand by storing heat in building materials and floors. Modern houses with floor heating proved far more flexible than older radiator-heated homes—a finding that matters as utilities worldwide seek to balance grids powered by variable renewables.EN

Researchers have developed a faster computational method for solving wave equations used in everything from seismic imaging to electromagnetic simulations. The breakthrough makes it possible to handle previously intractable real-world scenarios where wave properties vary unpredictably, potentially cutting computation time significantly for oil exploration, medical imaging, and telecommunications firms.EN

Researchers have developed a precise statistical technique to identify the beginning and end of COVID-19 waves—critical data for governments and businesses evaluating policy effectiveness. The analysis reveals waves last roughly 48 days, providing a quantifiable benchmark for assessing which interventions actually worked.EN

A new review shows machine learning algorithms are essential for managing the growing complexity of electric vehicle networks, from predicting battery performance to optimizing charging stations. As EV adoption accelerates, companies and utilities that master these AI techniques will gain competitive advantage in the shift to sustainable transportation.EN

Researchers have identified how to precisely control the composition of titanium boride coatings—used to harden aerospace components and cutting tools—by adjusting electrical pulse timing and intensity. The discovery could help manufacturers produce stronger, more durable coatings at scale, reducing costs and extending tool life in competitive industrial markets.EN

Researchers developed fiber-based sensors that track water temperature and pressure in real time for divers and send data to support ships above. The innovation targets a major safety gap: pressure and temperature changes cause most deep-sea diving fatalities, making reliable monitoring systems a potentially life-saving commercial opportunity for subsea industries.EN

Researchers have experimentally solved the microbunching instability, a technical barrier that has prevented free-electron lasers from achieving their full potential in the X-ray spectrum. The breakthrough could unlock new capabilities for materials science, drug discovery, and industrial inspection applications that depend on high-brightness X-ray sources.EN

Researchers have mapped how lithium-ion batteries produce dangerous gases during operation and identified sensor technologies that could detect these warning signs in real time. For manufacturers and EV makers, early detection could prevent recalls, fires, and device failures—a critical safety and liability issue as battery demand surges.EN

Researchers have developed a smarter way to distribute computing tasks across cloud data centers, reducing wasted processing power and energy use. The technique could help companies trim their cloud infrastructure bills while maintaining performance—a crucial advantage as enterprises face mounting pressure to control computing costs.EN

Researchers have cracked a decades-old computational bottleneck, delivering a method that solves complex physics equations 100 times faster than older approaches. The breakthrough matters: companies modeling heat transfer, fluid dynamics, or material behavior can now run simulations in hours instead of days, compressing timelines for product design and climate modeling.EN

Researchers have created detailed maps showing exactly where genetic recombination happens in plants—and found that the pattern varies significantly by species. The discovery could help crop breeders accelerate the development of hardier, higher-yielding varieties by targeting these genetic hotspots more precisely.EN