Tech & AI

Researchers have identified the optimal spacing between light sensors and emitters in wearable cardiovascular devices, potentially enabling thinner, cheaper monitors that work reliably on real skin. The findings could accelerate adoption of continuous heart monitoring for clinical care and consumer health tracking.EN

Researchers developed an AI-powered image recognition system that automatically identifies mechanical parts from customer photos, cutting out manual expert work. For manufacturing SMEs, this means faster service delivery and lower operational costs—critical advantages in industries where spare part delays directly cost customers money.EN

Researchers have identified why industrial automation systems often misconfigure their communication networks, leading to costly delays and equipment failures. The study provides practical rules for matching different types of factory traffic to proper network settings, a gap that has cost manufacturers real-time performance and interoperability across systems.EN

Researchers have pinpointed the root cause of degradation in spiro-OMeTAD, a widely-used material in high-efficiency perovskite solar cells: the chemical dopants added to boost performance are poisoning the device over time. The finding could unlock cheaper, longer-lasting solar technology critical for making renewables cost-competitive with fossil fuels.EN

A new study examines how play-based interactions help young children develop mathematical thinking. The findings could reshape early childhood education programs and influence policy around curriculum design, with potential long-term impacts on student achievement and workforce readiness.EN

Researchers have developed a machine learning system that predicts energy demand for parks and campuses with 29% greater accuracy than existing methods. The breakthrough could help facility managers reduce peak loads, lower utility costs, and improve grid stability as electrification increases demand.EN

A two-decade study of coal-dependent communities in India shows that transition plans miss informal workers and degraded ecosystems. The research argues justice requires listening to those most affected—a gap that could derail India's energy shift and create social instability.EN

A survey of 98 developers at Sweden's Transport Administration reveals that excessive meetings and task-switching—not talent gaps—are the biggest obstacles to faster software delivery. The finding offers a roadmap for government agencies and large organizations struggling to accelerate digital projects without expensive hiring.EN

Three major EU reports on AI competitiveness rely too heavily on technology-fix solutions while downplaying tensions with European values and ignoring how policies will affect smaller member states. The analysis warns policymakers that treating the EU as one entity could deepen internal divisions as the bloc races to compete with the US and China.EN

A new study shows that people expect AI assistants to handle negative emotions differently based on context and user gender—with some favoring calm reassurance and others preferring task-focused clarity. The findings challenge the industry's default approach of neutral responses, suggesting that smarter emotional calibration could improve user trust and reduce frustration.EN

Researchers argue that misinformation and polarization stem primarily from how social media platforms monetize user attention—not from free speech itself. The finding suggests policymakers should focus on structural reforms to platform economics rather than content moderation alone, reshaping a multibillion-dollar industry.EN

Researchers have developed a way to show users what data changes would alter AI forecasts—while honestly communicating prediction margins of error. The technique could help companies build trust in machine learning models used for high-stakes decisions like credit approval or medical diagnosis, where both accuracy and transparency matter.EN

A new collection of research from Europe's leading infrastructure security conference reveals how artificial intelligence is being deployed to defend power grids, transportation networks, and other essential services against cyber and physical threats. For utilities, governments, and infrastructure operators, the findings underscore both AI's promise as a security tool and the urgent need for new frameworks to manage AI-related risks in critical systems.EN

A researcher has challenged existing approaches to managing heat in electric mining truck batteries, questioning how thermal management systems perform in real-world conditions. The critique matters because thermal runaway in heavy-duty EV batteries remains a costly reliability risk—and mining operations depend on dependable equipment.EN

Researchers have created the first panoramic video dataset combining event cameras with standard HDR imaging to solve a persistent problem in virtual reality: motion blur and poor lighting in 360-degree content. The breakthrough could improve immersive experiences for VR platforms and headset makers struggling with video quality limitations.EN

Researchers have found that simulating wind sensations during virtual reality use significantly reduces motion sickness and boosts user immersion. The discovery could solve a major barrier to VR adoption in gaming, training, and workplace applications—where nausea currently limits session length and user willingness to engage.EN

A new international conference brings together researchers and practitioners focused on how artificial intelligence and information systems are reshaping business operations and digital strategy. The gathering signals growing demand from enterprises seeking to understand AI's practical applications beyond hype, from cybersecurity to healthcare to energy management.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 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 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 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

A new scoping review identifies how different software systems track and describe changes to their underlying models—a seemingly technical issue with real budget implications. Companies that adopt the wrong approach waste time recovering lost information about what actually changed, slowing updates to interconnected systems across their operations.EN

Researchers benchmarked seven large language models to automatically identify root causes in train control software test failures, a task currently requiring expensive manual labor. The work suggests AI could accelerate quality assurance in safety-critical industries where debugging delays increase costs and risk.EN

A new framework argues that digital twins—virtual replicas of physical systems—must continuously adapt as their real-world counterparts change, forcing companies to rethink how they build and maintain these expensive simulations. The insight matters because it could help manufacturers, infrastructure operators, and engineers avoid costly mismatches between their digital models and reality.EN

Researchers found that simulation-based safety tools let engineers spot thermal risks in electric vehicle batteries during early design stages, not after costly prototypes are built. The approach, tested in a university solar car project, could accelerate EV development and reduce design recalls.EN