Tech & AI

Researchers have discovered how to precisely control the electrical properties of scandium nitride—a promising material for converting heat into usable power—by embedding oxygen ions into the film. The technique could improve thermoelectric device efficiency for industrial waste-heat recovery and electronics cooling, a growing market as data centers and manufacturers seek ways to recapture energy.EN

Researchers used quantum physics simulations to explain why gold atoms segregate unevenly at grain boundaries in platinum-gold alloys—a finding that could improve the durability and performance of high-end alloys used in aerospace, electronics, and catalysts. Understanding this atomic behavior gives manufacturers a blueprint for designing stronger, more reliable materials.EN

Engineers have successfully scaled up a technology that turns sewage sludge, industrial waste, and other residues into combustible gas, achieving 77% energy conversion efficiency at a newly built demonstration plant. The breakthrough suggests waste-to-energy systems could soon become economically viable alternatives for industries seeking to reduce landfill costs while generating power.EN

Researchers have developed a faster way to simulate how soft tissues deform under stress—a crucial problem in medical device design and biomechanics. The method uses 40-70% less memory than traditional approaches, making complex simulations feasible on standard hardware and potentially accelerating product development cycles.EN

Researchers have proven that certain industrial control systems require exponentially more time and data to master than others. The findings establish hard limits on how fast autonomous systems can adapt—knowledge that matters for manufacturing, robotics, and energy grids where learning speed directly affects profitability and safety.EN

Researchers have identified precisely when a classical mathematical technique for analyzing complex signals breaks down and when it works reliably. The findings could refine how engineers filter data in telecommunications, audio processing, and other signal-heavy industries where accuracy at scale determines system performance.EN

Researchers have developed a cleaner mathematical approach to handling boundary conditions in computer simulations of physical systems, making them easier to implement and more reliable when data is incomplete. The method could accelerate engineering design, weather forecasting, and other fields that depend on solving equations across multiple interconnected domains.EN

Researchers examining internet infrastructure across Nordic and Baltic countries found that most DNS resolvers—the systems translating web addresses into usable data—have adopted modern security features, but a significant minority lack critical protections. The findings suggest uneven security deployment across the region's critical internet backbone, raising questions about data privacy and vulnerability to cyber attacks.EN

Researchers successfully operated a 1 MW gasifier using cashew shells—an agricultural byproduct—to produce syngas suitable for making synthetic fuels. The advance signals new economic potential for converting agricultural residues into chemicals and fuels, opening revenue streams for agribusiness while reducing waste disposal costs.EN

Computer scientists have developed faster methods to organize and simplify the logical rules that govern databases containing conflicting data types. The technique requires just one or two iterations to work effectively, making it practical for real-time systems where speed matters — from financial fraud detection to medical diagnostics.EN

Researchers have created a high-precision digital tank that models how waves and ocean currents interact with wave energy converters, a critical gap for commercializing renewable ocean power. The breakthrough could accelerate deployment timelines and reduce costly prototyping by allowing manufacturers to test devices virtually under varied sea conditions before building them.EN

Researchers have solved a major obstacle in rescue robotics: training AI systems without access to dangerous real-world data. By using photorealistic simulations, they created neural networks that can classify terrain and navigate disaster sites after learning only from synthetic environments—potentially accelerating deployment of autonomous rescue robots and reducing development costs.EN

Researchers have developed a deep learning system that spots unmanned aircraft by analyzing wireless signals rather than requiring dedicated radar or camera networks. The approach achieves over 99% accuracy in real-world tests and could dramatically cut costs for airports, urban authorities, and enterprises managing airspace security.EN

Researchers have developed a power allocation method that lets wireless networks maximize overall throughput while ensuring no single user gets starved of bandwidth. The technique matters for telecom operators building equitable 5G/6G systems and for IoT deployments where fairness affects service quality guarantees.EN

Computer scientists have created a unified framework for understanding different mathematical models used to build modern programming languages and type systems. The work addresses a 30-year gap in the field by showing how competing approaches actually fit together—a finding that could simplify how language designers build more reliable software tools.EN

Researchers have proven that three commonly used computational methods can reliably solve complex flow problems if properly calibrated—a finding that could accelerate everything from climate forecasting to financial modeling. The work removes a major barrier holding back adoption of faster, cheaper simulation techniques in industries that depend on real-time predictions.EN

Researchers have developed a method to identify active IoT devices on wireless networks without using traditional pilot signals, reducing network overhead and enabling denser device deployments. The approach works by analyzing signal angles rather than full channel data, offering a scalable solution for operators managing millions of connected devices.EN

Researchers have identified how defects in low-cost radio hardware degrade network performance in next-generation 5G and 6G systems—and designed fixes that recover lost efficiency. The finding matters because telecom operators planning dense networks of budget access points now have a roadmap to avoid costly performance penalties.EN

Researchers have solved a longstanding problem in 5G and beyond networks: how to process signals efficiently when data travels through multiple layers of a communications system. The breakthrough reduces the amount of data sent across network infrastructure by up to 50%, lowering operational costs for carriers while maintaining call quality and data speeds.EN

Computer scientists have developed a faster way for networks of machines to solve complex optimization problems by combining two established techniques. The approach could accelerate decision-making in cloud computing, robotics, and large-scale data analysis—domains where companies currently lose efficiency coordinating calculations across multiple servers.EN

A new study tracking how phosphorus behaves when grass and brewery waste are burned for energy shows that most of the nutrient stays in coarse ash rather than polluting the air. The finding matters for companies designing biomass power plants and waste-to-energy facilities seeking to improve ash management and reduce environmental compliance costs.EN

Researchers are developing computational methods to help artificial intelligence understand spatial relationships and solve real-world navigation problems. The work matters to businesses building autonomous systems, robotics, and location-based services—making AI smarter about how it interacts with physical space.EN

Researchers have developed a system that helps AI models adapt when data patterns change over time across decentralized networks—a common real-world problem that current systems struggle to handle. The breakthrough matters for enterprises deploying machine learning across multiple locations, from healthcare to finance, where both data privacy and model accuracy depend on handling non-stationary environments.EN

Researchers have exposed a critical blind spot in AI video understanding systems: they work primarily in English and struggle with non-English languages and cultural content. A new test covering 14 languages and culturally specific scenarios reveals how current AI tools may exclude billions of potential users and markets.EN

Researchers have developed a task allocation algorithm that improves how multiple satellites coordinate communication routing in orbit. The advance could reduce latency and boost reliability for satellite networks serving everything from remote connectivity to autonomous vehicles—a critical capability as the industry prepares for mega-constellations with thousands of orbiting nodes.EN