Tech & AI

Researchers have developed a way to strengthen plywood adhesives by adding micro-fibrillated cellulose to lignin-based formulations, reducing reliance on petroleum-derived chemicals. The finding could help plywood manufacturers lower costs and environmental impact while maintaining structural quality.EN

Researchers have developed a new prediction system that forecasts human movement patterns far longer and more accurately than existing methods, critical for autonomous vehicles and service robots operating around people. The breakthrough could significantly reduce collision risks and improve safety in shared human-robot spaces.EN

Researchers have developed a radar-based navigation system that matches lidar's accuracy—a breakthrough that could slash equipment costs for autonomous vehicles and warehouse robots. The technique works across diverse environments without retuning, making it practical for commercial deployment at scale.EN

Researchers have developed a fuzzy clustering method that can align multiple medical images and automatically check its own work for accuracy. The technique could streamline clinical workflows and reduce manual quality control costs for hospitals and imaging companies processing large datasets.EN

A team has fixed a long-standing gap in automated theorem proving by enabling specialized math solvers to handle floating-point arithmetic—the backbone of scientific computing. The upgrade significantly boosts success rates in formal verification, a critical tool for ensuring software reliability in safety-critical applications like aviation and autonomous systems.EN

Researchers tested whether service robots can safely mask patients' faces while working in doctor's offices. The technology works in controlled settings but fails when patients move, lighting changes, or they're partially hidden—raising questions about robot deployment in hospitals before the privacy problem is solved.EN

As India's oldest coal mines shut down for economic reasons, a new study reveals that closure policies fail to protect dependent communities long-term. Researchers found that voluntary guidelines produce inconsistent results—some sites become forests or tourism parks while workers get only temporary jobs, raising questions about whether India can manage large-scale industrial transitions without stronger regulation.EN

A new study shows that stone matrix asphalt overlays last significantly longer than conventional hot mix asphalt and remain cost-effective despite higher upfront prices. For transportation departments managing tight budgets, the finding suggests that choosing the right pavement type based on existing road condition can slash maintenance spending over decades.EN

Researchers have developed a technique that prevents AI models from forgetting previously learned information when being adapted for new tasks—a common and costly problem in machine learning. The approach could reduce retraining time and expenses for companies deploying pre-built AI systems across multiple applications.EN

Researchers have developed a machine learning system that predicts methane output with greater accuracy by analyzing complex production data at multiple time scales simultaneously. The advance could help energy companies optimize operations, reduce waste, and strengthen supply chain resilience as global demand for methane as a chemical feedstock and fuel grows.EN

Researchers have discovered that the physical shape of antenna arrays dramatically affects how much data wireless networks can transmit at close range—a critical capability for future 6G systems. The finding could let telecom operators squeeze more simultaneous users onto the same spectrum by redesigning array geometry, potentially worth billions in network efficiency gains.EN

Researchers used molecular modeling to map exactly how lignin—a byproduct of paper and biofuel production—breaks down into useful gases when heated in CO2 or water vapor. The findings could unlock a cheap way to convert billions of tons of industrial waste into synthetic fuels and chemicals, reducing landfill burden while creating new revenue streams for manufacturers.EN

Researchers have identified a class of alloys that dramatically improve the efficiency of water-splitting catalysts—a critical step in producing clean hydrogen fuel. The breakthrough could accelerate commercial viability of hydrogen as an energy source and reduce capital costs for renewable energy infrastructure.EN

A new analysis reveals that IKEA's store network expansion in Sweden between 2004 and 2016 reduced average travel distances to stores, but could have done better. Researchers found the company could have achieved the same environmental gains with fewer, better-placed locations—a lesson for any retailer balancing growth with sustainability commitments.EN

Scientists have mapped how the atomic structure of zinc oxide thin films affects their ability to convert sunlight into electricity, achieving a 3.39 electron-volt bandgap ideal for solar applications. The findings—validated through both lab experiments and computer modeling—could accelerate development of cheaper, more efficient solar panels for manufacturers competing in the renewable energy market.EN

Researchers have developed a framework to determine how roots of complex equations shift as their parameters change—a problem relevant to cryptography, control systems, and AI model stability. The work provides practical tools for engineers and analysts who need to predict system behavior without solving equations from scratch.EN

Researchers have developed a technique that makes electricity price predictions significantly more accurate by reconciling forecasts across different time blocks simultaneously. The method works with any AI model and costs almost nothing extra to run, making it immediately practical for power traders and grid operators managing billions in daily transactions.EN

Researchers have solved a decades-old mathematical puzzle about how waves scatter across complex surfaces—work that could accelerate quantum algorithms and improve wireless systems. The breakthrough provides explicit formulas for predicting wave behavior, offering practical tools for engineers designing next-generation communication networks and quantum technologies.EN

Researchers developed an AI system that precisely segments leg muscles in MRI images, enabling doctors to quickly measure muscle volume without manual analysis. The breakthrough could transform how clinicians assess rehabilitation progress in stroke patients and other conditions affecting muscle loss, reducing assessment time while improving accuracy for clinical trials and personalized treatment planning.EN

Researchers have created ORCA, a four-principle framework that guides teams through the messy process of tackling complex societal challenges where problems and solutions evolve together. The approach, tested on a blockchain water-exchange project, offers business and government leaders a practical roadmap for innovation in uncertain, high-stakes environments.EN

Researchers have developed a method to predict how systems fail when components depend on each other in sequence—a problem that traditional reliability models miss. For industries managing critical infrastructure, this could mean better predictions of cascading failures and smarter maintenance strategies that prevent costly downtime.EN

Researchers have developed an AI system that enhances low-light images while preserving detail and controlling noise—a capability critical for autonomous vehicles, surveillance, and mobile devices. By embedding physics-based noise detection into a machine learning model, the framework achieves better results faster than existing methods, potentially unlocking new applications in robotics and visual inspection.EN

Researchers have developed FlexSpec, a system that decouples mobile AI models from cloud versions, eliminating the need to constantly resync or retrain edge devices when the main model improves. For companies deploying AI at the network's edge—from autonomous vehicles to factory sensors—this cuts communication costs and speeds up inference by 30-50%, making real-time AI deployments more practical and scalable.EN

Researchers found that basic cognitive processes — not specialized executive abilities — explain most similarities in how people inhibit impulses, shift tasks, and update information. The finding could reshape how companies and institutions design cognitive assessments, employee training programs, and AI systems that attempt to replicate human decision-making.EN

Researchers have developed a new approach to recognize individuals based on body shape and gait rather than clothing, solving a long-standing challenge in surveillance and security systems. The breakthrough could improve airport screening, law enforcement investigations, and loss prevention—while raising fresh privacy concerns for regulators and security teams.EN