Tech & AI

Scientists have validated waste materials like broken roof tiles and cement dust as viable thermal storage media, stable at 800°C through hundreds of heating cycles. This breakthrough could dramatically reduce costs for grid-scale energy storage systems, making it cheaper to store excess solar and wind power for later use.EN

Researchers have shown that AI-based time-series forecasting significantly improves predictions of power transformer temperatures compared to standards used since the 1980s. Better temperature forecasting could reduce unexpected transformer failures and extend equipment lifespan—cutting maintenance costs for utilities managing aging electrical grids.EN

Researchers have created a mathematical framework that predicts how people's willingness to get tested and isolate influences infectious disease outcomes. The model reveals critical feedback loops between behavior and disease dynamics that could help policymakers design more effective outbreak response strategies and predict which interventions actually work.EN

Scientists have demonstrated a working microwave photon detector that captures nearly 70% of the faintest quantum signals—a major efficiency gain for quantum computing and communication systems. The breakthrough uses semiconductor quantum dots coupled to superconducting cavities, offering a scalable platform that could accelerate commercial quantum technology development.EN

Scientists have developed a mathematical method to drastically reduce the complexity of large networked systems—from power grids to neural networks—while preserving exact accuracy. The breakthrough could accelerate simulations and analysis of systems too large to model directly, offering immediate value for infrastructure operators and AI developers managing scale.EN

Researchers have developed a method to control how electrical charges move through semiconductor interfaces using mechanical stress, enabling pressure sensors far more sensitive than existing commercial technology. The breakthrough could reshape sensing applications in communications, industrial monitoring, and consumer devices where precision pressure detection drives performance.EN

Researchers have identified two overlooked failure mechanisms in a promising renewable battery design, using real-time electrical monitoring to catch performance killers before they crash system lifespans. The findings could reshape how energy companies predict and extend the reliability of next-generation grid batteries meant to stabilize renewable power.EN

Researchers have created flexible structures that move and adapt on their own without external control, by harnessing an unusual physics phenomenon called exceptional points. The breakthrough could enable more autonomous robots, smart materials, and mechanical systems that respond to their environment without constant human input or tethering.EN

Researchers have cracked how to read muscle electrical signals to detect when people are straining, offering a faster and cheaper alternative to measuring energy burn for optimizing robotic limb assistance. The finding could accelerate deployment of exoskeletons in warehouses and healthcare by enabling real-time, personalized device tuning without lab equipment.EN

Researchers have cracked a manufacturing bottleneck in perovskite solar cells using vacuum deposition—a scalable, solvent-free factory method. The breakthrough delivers certified 18.35% efficiency for single-layer devices and 27.2% for perovskite-silicon tandems, with cells retaining 80% performance after over 1,000 hours of stress testing, positioning the technology closer to commercial viability.EN

Researchers found that adding video to train language models helps only certain words, not all vocabulary equally. The discovery suggests companies investing in multimodal AI need to rethink training strategies based on what kinds of information models actually need to learn effectively.EN

Researchers have identified a critical flaw in next-generation wireless systems: conventional antenna designs fail catastrophically when obstructed by people or objects. A new radio-over-fiber architecture distributes antennas across larger spaces, dramatically reducing blockage vulnerability and cutting deployment costs for ultra-high-speed 6G networks.EN

Researchers have developed a method that extracts meaningful patterns from complex, noisy datasets—from brain scans to particle simulations—using topological data analysis integrated into visual exploration tools. The breakthrough could accelerate discoveries in neuroscience and materials engineering by helping scientists spot trends faster and compare results more reliably across studies.EN

A randomized controlled trial shows that experiencing virtual underwater environments while physically immersed in water significantly increases people's commitment to ocean protection—and the effect persists a month later. For conservation organizations, tourism operators, and policymakers seeking cost-effective ways to drive pro-environmental behavior, the findings suggest immersive VR could transform how we build public support for marine ecosystem protection.EN

Researchers have identified a new type of electron crystal in twisted semiconductor bilayers that behaves counterintuitively—its quantum properties cancel out despite occupying multiple energy bands. The discovery could reshape how engineers design next-generation quantum computing materials and topological devices, where controlling electron behavior at the atomic scale is critical.EN

Research shows that worn wheels and widened track gauge are the primary culprits behind accelerated damage on the inner rails of tight curves in heavy-haul operations. The findings could help rail operators reduce costly maintenance downtime by targeting preventive measures more precisely at high-risk track sections.EN

A new study of Kenya and Liberia shows that rolling out digital health tools without fixing underlying connectivity and affordability problems leaves rural populations behind. The finding matters to health tech investors and policymakers: expensive apps and platforms won't work without cheaper data, reliable electricity, and training—a costly reality most implementations ignore.EN

A new study shows that machine learning models trained in controlled laboratory settings often break down when deployed in actual industrial environments. Transfer learning techniques can close this gap by up to 30%, potentially unlocking AI-powered predictive maintenance across manufacturing and critical infrastructure industries where equipment failures are costly.EN

Researchers have mapped the modular organization of mammalian spinal locomotor networks and their real-time dynamics, revealing how the nervous system coordinates complex movement. The findings could inform development of neural interfaces, prosthetics, and treatments for paralysis and movement disorders—opening commercial opportunities in neurotechnology and regenerative medicine.EN

Researchers have successfully grown and printed astrocytes—support cells critical to brain function—using engineered hydrogel scaffolds. The breakthrough could accelerate drug testing, reduce reliance on animal models, and open new avenues for treating neurological diseases, potentially reshaping how biotech and pharmaceutical companies develop therapies.EN

Researchers have validated an artificial intelligence system that analyzes mammogram images to forecast a woman's breast cancer risk over the next several years with sustained accuracy. The finding could reshape how hospitals screen patients and allocate resources, potentially preventing thousands of cancers through earlier intervention.EN

Researchers developed a screening method that detects how accidentally-formed metal nanoparticles harm human cells at the molecular level. The approach could reshape how manufacturers monitor product safety and regulatory compliance, potentially preventing costly recalls and liability claims.EN

Researchers have expanded the mathematical foundations underlying quantum computing by discovering how to construct new types of quantum groups—abstract structures that describe quantum systems. The work bridges previously disconnected areas of quantum mathematics, potentially opening pathways for engineers to design more versatile quantum hardware and algorithms.EN

Researchers have cracked a long-standing problem in perovskite solar cells—uneven material distribution that degrades performance. By using a seed-layer technique to control how atoms arrange during manufacturing, they've achieved certified efficiency gains and created flexible solar panels. The breakthrough could accelerate commercialization of lighter, cheaper alternatives to silicon cells.EN

Scientists have created a computational framework that can track and verify how human emotions change in response to observable actions, using formal logic and psychology theory. The work could shape how AI systems that interact with people—from customer service bots to mental health apps—are designed and audited for unintended emotional influence.EN