Tech & AI

Researchers have upgraded a popular artificial intelligence technique to prevent it from getting stuck on inferior answers—a chronic problem that wastes computing resources. The enhancement could make optimization software faster and more reliable for industries relying on complex calculations, from logistics to engineering design.EN

Researchers have developed a cleaner way to analyze a type of random mathematical structure used in computer science and probability theory. The new approach makes calculations more transparent and accessible, potentially enabling faster development of algorithms that rely on tree-based data structures in machine learning and computational biology.EN

A global survey of 230 researchers reveals a stark gap: while institutions worldwide have adopted open science policies, almost none provide formal recognition or career incentives for researchers who participate. The disconnect threatens adoption rates and could undermine billions in research investments unless institutions align rewards with stated values.EN

A new study tracking smart grid development in an Arctic city reveals how digital networks are fundamentally reshaping the relationship between utilities and customers. As energy systems evolve from one-way transmission to two-way platforms, residents shift from passive consumers to active participants—a transformation with major implications for grid stability and energy transition investments.EN

Researchers have developed a hybrid optimization technique that combines strengths of two competing algorithms to solve difficult computational problems more efficiently. The breakthrough could accelerate machine learning training, industrial design, and other applications where finding the best solution among millions of options is critical to business performance.EN

Researchers created LYSTO, a benchmark dataset that enabled AI systems to match pathologist accuracy in counting lymphocytes—key immune cells—in cancer biopsies. The work signals that machine learning can reliably automate labor-intensive microscopy tasks, potentially accelerating cancer diagnosis and reducing costs for hospitals and diagnostic labs.EN

Researchers have solved a decades-old puzzle about how networks organize themselves, discovering that the answer depends entirely on the network's structure. The findings could help companies optimize everything from supply chains to social networks by predicting how interconnected systems behave.EN

Researchers have developed a faster, more accurate method for compressing 3D product shape data used in computer-aided design software. The breakthrough could reduce file sizes and transfer times for manufacturers managing complex product designs, potentially cutting costs in industries from automotive to aerospace.EN

A trucking fleet cut dangerous speeding violations substantially by installing in-vehicle monitoring systems that provide drivers immediate feedback on their behavior. The effect was strongest for the most reckless drivers, suggesting targeted safety programs could reduce the roughly one-third of large truck crashes tied to speeding—saving lives and liability costs.EN

A new analysis of Rio de Janeiro's waste streams shows the state could generate enough biogas to replace up to 22% of its natural gas demand and produce enough fertilizer to meet regional agricultural needs. The finding suggests a viable path for reducing fossil fuel dependence while tackling emissions from livestock farming and waste management.EN

Researchers developed an assessment tool that identifies exactly which foundational skills—not just advanced concepts—cause computer science students to struggle. The findings could help universities redesign programming curricula and improve student retention in tech talent pipelines.EN

Researchers have created design guidelines for machine learning devices built on biodegradable circuit boards made from plant-based materials, addressing the growing e-waste crisis from discarded electronics. The work suggests a path toward sustainable AI hardware that can degrade after use rather than accumulating in landfills—a shift that could reshape how tech companies approach device manufacturing and end-of-life planning.EN

Researchers have solved a long-standing problem in measuring how chemical reactions unfold on material surfaces, opening faster paths to develop new semiconductors, catalysts, and nanotech products. The method cuts through ambiguity in existing measurements, letting companies predict reaction behavior with greater accuracy and fewer costly experiments.EN

Researchers have replaced cadmium in antimony selenide solar cells with titanium oxide, eliminating a toxic material while achieving record 8.12% efficiency. The breakthrough could accelerate commercial deployment of cheaper, more sustainable photovoltaic technology that requires fewer processing steps than conventional silicon panels.EN

Researchers have developed a faster way to design antiviral drugs by starting with tiny molecular fragments and building them up to hit a critical COVID-19 target. The approach produced a compound that binds effectively to the virus's main protease, potentially accelerating the discovery of next-generation treatments for coronavirus variants.EN

Researchers have engineered a photodetector using a two-dimensional material that responds to light wavelengths spanning the entire visible spectrum—from ultraviolet through red. The device maintained stable performance through 15,000 cycles, suggesting potential applications in imaging sensors, optical communications, and autonomous systems where broad-spectrum light detection is required.EN

Researchers have developed a faster way to detect correlations in multi-factor experiments—a common challenge in manufacturing, agriculture, and quality control. The approach could help companies spot efficiency problems or defects earlier, potentially saving millions in production costs and reducing time spent on data analysis.EN

Researchers have designed molecules that emit light more efficiently in solid form, a key requirement for next-generation displays. The advance addresses a persistent engineering challenge in organic light-emitting diodes—getting materials to glow brightly when packed into actual devices rather than in laboratory solutions.EN

Researchers have created a molecule that simultaneously inhibits two cancer-driving proteins, showing strong activity against breast cancer cells in lab tests. The dual-target approach could overcome drug resistance and improve outcomes for patients with hard-to-treat tumors, potentially opening a new commercial category for cancer therapeutics.EN

A 24-year Finnish study tracking workers exposed to industrial vapors, gases, dusts, and fumes found they face 2.1 times higher risk of fatal heart attacks and 1.7 times higher risk of dementia-related death. The findings could reshape occupational safety standards and workplace health monitoring across manufacturing, construction, and chemical industries.EN

Researchers analyzing nearly 150,000 residents found that proximity to public squares and building heights significantly influence perceived safety—suggesting urban planners can reduce neighborhood disparities through design choices rather than expensive individual interventions. The findings offer cities a cost-effective tool for addressing inequality and improving quality of life.EN

Researchers have cracked a long-standing mathematical problem that could unlock better computer simulations of how liquids and gases interact with deformable solids—critical for engineering everything from aircraft wings to artificial organs. The breakthrough proves these complex interactions can be reliably modeled over long time periods, removing a major barrier to digital design and testing.EN

Researchers used a novel statistical method to show that differences in diabetes and high cholesterol explain only part of why men suffer more heart attacks than women—socioeconomic status plays a major role. The findings suggest healthcare systems and insurers need to account for social factors, not just medical risk factors, when predicting cardiovascular disease risk across populations.EN

Chinese researchers used 2.7 billion particle collisions to measure how often a subatomic particle decays in a specific way, pinpointing a value that had eluded scientists for years. The finding validates detector technology and computational methods used across high-energy physics experiments worth billions in global research investment.EN

Researchers have developed a faster computational method to measure how light scatters and absorbs in industrial materials like titanium dioxide and metallic suspensions. The technique cuts analysis time and cost compared to existing approaches, potentially accelerating product development cycles in manufacturing sectors that rely on precise optical property characterization.EN