A new AI system, PI-GAE, predicts aerodynamic performance of shapes, simplifying the design process for engineers. The system accurately reconstructs shapes and predicts performance, with coefficients of determination ranging from 0.946 to 0.992 across four tests.
Researchers at Dartmouth College found that pink noise, a conversational recipe balancing familiarity and unexpectedness, tends to be rated as more enjoyable. The team's study suggests that injecting randomness and switching topics can help revitalize conversations and improve human communication.
Researchers at the University of Tartu have developed a method to create detailed 3D models of plants using consumer-grade cameras, allowing for the preservation of living collections. The technique, combining photogrammetry and 3D Gaussian Splatting, enables the capture of plant growth over time and creates interactive virtual walkthr...
Researchers at OIST have pioneered a forward path to complex curved structures, characterizing a new class of shapes: ortho-arc curves (OACs). They enable simpler, faster access to curved shapes with controllable properties such as symmetry, chirality, or topology.
Researchers have developed a unified AI architecture, WorldParticle, that can simulate multiple physical phenomena, from sliding sand to folding proteins, with one unified artificial intelligence architecture. The system represents every material as tiny particles and predicts where those particles will move one step at a time, adaptin...
Researchers from Politecnico di Milano developed a lithium-free, non-flammable battery with a fluoride ion membrane that lasts twice as long as traditional lithium-ion batteries. The membrane enables fluoride ion transport at room temperature, reducing the risk of fire and increasing energy density.
Five new research hubs will develop tools to detect early signs of illness, such as dementia and traumatic brain injuries, and prevent them. The hubs will bring together experts from across the UK to create innovative solutions, including AI-powered brain health clinics and computer models to predict brain injury risk.
A study published in Nature Machine Intelligence found that many advanced AI models, including those used in language processing and decision-making, struggle to counter stereotypical assumptions. The researchers used the Implicit Association Test to measure the computational effort required to process biased information, and found tha...
Researchers develop Physics-based Learning for Imaging Planets (FLIP) to remove starlight from telescope data, revealing exoplanets previously hidden. The method improves detection sensitivity, enabling scientists to study more planets and uncover previously missed worlds.
A unified model classifies electrochemical, mechanical, and thermal battery models, charting a path toward safer energy storage. The review gathers over a decade of modeling work, spanning scales from single particles to complete systems.
Researchers created the most accurate asteroid deflection model to date, considering realistic asteroid shapes and high-fidelity math. The study found significant variations in deflection velocities and spin rates depending on the asteroid's shape and target area, highlighting the importance of precise aiming in a nuclear blast.
Researchers have developed a new mean-field model that incorporates ion exchange mechanisms, allowing for more realistic simulations of brain activity. This model can capture the complex interactions between neurons and their surroundings, which are crucial for understanding neurological disorders such as epilepsy.
The new simulation, Lumina, captures the emergence of two huge transformations in cosmic history: hydrogen re-ionization and helium re-ionization. Hydrogen re-ionization was likely driven by light from early galaxies and stars, while helium re-ionization was fueled by energetic early supermassive black holes.
A new AI-based computer model has been developed to predict treatment outcomes for patients with Triple-Negative Breast Cancer, a hard-to-treat cancer. The model uses symbolic AI and accurately predicted treatment effectiveness in 74% of cases, suggesting new therapeutic strategies.
Geometry-based kinetic simulations reveal conditions for improving DNA-nano/microparticle motor speed, run length, and unidirectionality. Large particles can increase run length and unidirectionality due to increased DNA/RNA duplex formation, while small particles retain room for further acceleration.
Researchers developed a machine learning approach to reconstruct missing precipitation patterns in global rainfall maps, using a 3D U-Net and 3D condition encoder to capture space and time variations. The model achieved high prediction accuracy and produced consistent rainfall patterns across locations and time.
Researchers recreated the Imperial Woodpecker from a museum specimen and 1956 film footage, providing a detailed, animated 3D model. The model is freely available and can be used for scientific research and public awareness, offering insights into the species' behavior and cultural significance.
Researchers developed an aging-specific screening platform that identified SRSF1 as a gene with multifaceted rejuvenation potential, reducing hallmarks of aging and extending lifespan in mice by 14%. The study provides a promising avenue for devising clinically useful de-aging strategies.
Artificial intelligence can match human doctors in certain tasks, but human doctors are essential for making complex decisions about patient care. AI may identify a patient's medical condition, but human doctors must consider individual circumstances and make choices that work in the patient's best interest.
Researchers developed a method to adapt adult-trained AI models for pediatric use, addressing critical gaps in medical AI for children. The method, trained on adult data, outperformed adult models trained on pediatric data from a single hospital at 29 of 30 sites, even when tested on new images.
A new path-tracing method simulates event camera data from virtual 3D scenes efficiently, reducing computational costs for applications like autonomous driving and robotics. The method uses physically-based path tracing and adaptive temporal search to precisely determine event timings, enabling more efficient training-data generation.
SourceChiba University·JournalIEEE Transactions on Visualization and Computer Graphics·TypeComputational simulation/modeling·DateOct 1, 2026
Researchers are working on a four-year project to develop a personalized model that uses detailed blood flow information to detect obstructions in the aorta. The project aims to improve prenatal testing so that doctors can spot problems in the aorta before a baby is born and recommend treatments.
UVA Engineering AI researcher Chen Chen is developing trustworthy AI systems to forecast infectious disease epidemics, combining multiple data sources like reported cases, wastewater surveillance, and human travel data. Her approach aims to predict where diseases will spread and help public health officials respond effectively.
Researchers at the Stowers Institute used AlphaFold2 and evolutionary data to predict protein structures in aphids, which were previously inaccessible to AI. The study reveals a common architectural plan among 2,400 BICYCLE proteins, showcasing the evolution's role in helping AI predict protein structures.
Researchers from Tokyo Metropolitan University used evolutionary algorithms to identify optimal shapes for ultra-thin, bio-inspired corrugated airfoils. They discovered designs that minimized drag or maximized lift, with corrugations near the leading-edge reducing frictional drag and convex shapes near the trailing edge increasing lift.
Researchers are using AI to model complex turbulent flows, including wind, aerosols, and combustion, to quantify uncertainty in foundation models. The project aims to create a physics-constrained AI foundation model for complex flows important to energy systems.
A new AI tool has been developed to predict treatment and survival outcomes in patients with advanced non-small cell lung cancer treated with immunotherapy. The tool consistently outperformed standard clinical biomarkers, achieving high accuracy scores in predicting survival outcomes.
Researchers will develop AI-powered tools to examine cascading wildfire impacts, including flash floods, debris flows, and infrastructure disruptions. The project aims to bolster public safety and community resilience by combining AI, infrastructure modeling, and social science.
Team Kinetiz from NTU Singapore, K2, and Alp Autonomy won the world's first multi-car autonomous race in Europe, taking the lead at Imola Circuit. The team demonstrated exceptional autonomous driving skills, showcasing the potential of AI in motorsport.
Researchers at Chalmers University of Technology have developed a new method for performing advanced quantum operations significantly faster and more efficiently. This breakthrough addresses a well-known bottleneck in quantum computing and paves the way for fault-tolerant quantum computing.
A new learning mechanism uses natural variability in neural activity to understand how synapses adapt and improve the learning capabilities of brain-inspired devices. The mechanism, called Spike-based Alignment Learning, solves the weight transport problem and matches the performance of existing approaches without unrealistic assumptions.
The new MSc Computer Science with Artificial Intelligence programme equips learners with advanced technical knowledge and practical skills in AI. The programme combines Heriot-Watt's academic expertise with flexible online delivery, enabling learners worldwide to study alongside work and other commitments.
Researchers found that AI-generated images can improve species identification and biodiversity monitoring when real images are limited. However, the synthetic images were less effective than real images overall, highlighting the importance of community science and real-world observations.
Researchers developed a physics-based AI approach to predict global-scale carbon cycling in ocean sediments, resolving a long-standing challenge in climate science. The study reveals key findings on dissolved organic carbon behavior, including 11% of particulate organic carbon returning to seawater.
Researchers from SUTD develop HieraScaffold, an AI framework that generates large-scale 4D LiDAR scenes more efficiently and coherently. The framework captures both static structures and moving objects, improving the realism and accuracy of autonomous systems.
Researchers at ETH Zurich have created a 3D-printed model of a rock avalanche to study the movement of mixtures of water, ice, and rock. The model, which is 1:577 in scale, was used to test various scenarios and measure parameters such as depth of runoff and impact dynamics.
A machine-learning tool built from Swedish national health registry data can predict hip fracture risk with high accuracy and identify individuals at high risk without in-person assessment. The tool performed nearly seven times better than current screening methods in identifying at-risk individuals.
SourcePLOS·JournalPLOS Medicine·TypeComputational simulation/modeling·DateAug 27, 2026
Researchers developed a novel divide-and-conquer approach for model checking linear temporal properties, called DCA2MC, to address state-space explosion and long verification times. The approach divides the original model checking problem into smaller, independent tasks, reducing memory consumption and verification time.
Engineers at Texas A&M University, NASA, and Purdue University create algorithms for managing spacecraft traffic around Gateway, a lunar spaceport. The system balances fuel efficiency and operational demands to reduce the risk of collisions during space missions.
MIT engineers develop a tool that generates plausible extreme events and worst-case scenarios without relying on extreme data, enabling planners to prepare for unprecedented scenarios. The algorithm takes a statistical approach to learn from available data, excluding implausible weather scenarios, and projects how extreme events might ...
The ST-NUS HELIX Corporate Lab aims to develop new generative and embodied AI use cases at the edge through system-to-silicon innovation, reducing energy consumption and improving performance. Researchers will focus on memory-centric architecture, innovative in-memory computing, and scalable compute-and-memory systems.
The Skala AI model, developed by Microsoft Research AI for Science, is now available through the CP2K software ecosystem. CASUS and Microsoft Research collaborated to integrate Skala into CP2K, enabling more accurate quantum mechanical simulations of larger molecular systems. The collaboration aims to improve the accuracy and efficienc...
A new study models a future net-zero European power system and tests it against 80 years of historical weather data to understand how it would handle stress with periods of low wind and solar generation, combined with high demand. The study concludes that the risk is greatest during winter when cold and wind-still conditions persist, h...
A new AI model, EarlyDetect, can detect precursor signals of active region emergence in the Sun's acoustic activity and magnetic field, forecasting solar eruptions nearly nine hours in advance. This technology has the potential to allow satellite communications companies or power grid companies to prepare for solar storms.
Researchers developed Latent Seal, a watermarking framework that integrates image watermarks into latent diffusion models. The approach helps identify AI-generated content and support copyright verification without degrading image quality. Tests show the method remains accurate after common edits and distortions.
Researchers built a mathematical model to understand collective movement and rest in ant colonies. The findings show that a single 'first mover' ant activates a group, triggering a wave of coordinated movement involving hundreds or thousands of workers.
Researchers developed a method to predict how different types of breast cancer will respond to treatment using lab-grown mini tumors. The organoids mimicked a tumor's response to treatment and identified candidate combination therapies for cancers that don't respond to standard treatment.
Assistant Professor Yingxue Zhang's project aims to develop urban AI models that can efficiently process vast amounts of human-generated data to optimize commute times, traffic safety, and more. The model will utilize offline reinforcement learning to tackle spatial-temporal dynamics in urban life.
A simulation study from the University of Illinois suggests that winter canola can increase overall productivity by 18% while maintaining stable greenhouse gas intensity. The diversified system also improves net ecosystem carbon balance, with benefits increasing over time.
New York's Empire AI Beta has officially launched, providing world-class AI computing power to researchers across the state. The initiative has served as a model for the federal National Science Foundation's State and Regional AI Infrastructure Hubs, which aim to build out regional AI research infrastructure and shared research capacity.
Research at AIST, Institute of Science Tokyo, and Ritsumeikan University uncovers the impact of histone acetylation site location on liquid–liquid phase separation in gene regulation. This study provides new perspectives for therapeutic development targeting aberrant gene expression.
A research team from HKU School of Computing and Data Science has developed ClairS, a novel deep-learning algorithm that improves the detection of cancer mutations. Tested on breast, lung, and melanoma cell-line datasets, ClairS demonstrates exceptional accuracy across various cancer types and sequencing conditions.
The Smart QS Hackathon 2026 aims to explore innovative applications of Artificial Intelligence (AI) in Quantity Surveying (QS) through cross-sector collaboration, facilitating the intelligent transformation of the construction industry. Participants will have access to AI training sessions and workshops, as well as Cyberport's computin...
Researchers found that the Agulhas Current supplies heat and moisture that strengthen weather systems, causing severe floods in the Western Cape. In contrast, the Benguela Upwelling System has a minor role in intensifying rainfall.
Researchers developed three functional components for photonic microchips using inverse design algorithms. The new components are up to 500 times smaller and more efficient than traditional designs.
Researchers developed an inverse-design framework to optimize magnonic crystal design, identifying unconventional lattice structures with large band gaps. The approach enables the exploration of previously unexplored material systems and device dimensions, paving the way for high-speed spin-wave computing and energy-efficient devices
USC researchers have been selected for the U.S. Department of Energy's Genesis Mission to harness artificial intelligence for scientific discovery and innovation. Two projects led by USC will explore ways to develop faster and more energy-efficient computing hardware and better understand the natural concentration of critical minerals.
Researchers at Duke University have developed a method to systematically develop novel probiotic and prebiotic combinations to maintain gut health and treat gastrointestinal diseases. The approach uses machine learning and automation to explore complex interactions between microbes, nutritional sources, and the environment.
A team of researchers at Harvard and Max Planck Institute have developed three new functional components for photonic microchips using an inverse design algorithm. The compact designs are about 500 times smaller than conventional designs and offer a path toward higher-performance integrated light technologies.
Two Lehigh University AI projects have been selected for funding from the Department of Energy's Genesis Mission. The RIVER-AI project will improve flood- and water-level prediction, while the REACT project aims to accelerate reactor-scale fusion energy by developing an AI-enabled digital twin. These awards strengthen Lehigh University...