Articles tagged with Algorithms
The University of New Mexico-led project, Cognitive Autonomy for Human CPS: Turning Novices into Experts, seeks to develop a framework for improving autonomous technology's adaptability to real-life human behavior. The team will investigate ways to anticipate and mitigate common human-automation interaction problems.
Researchers at Duke University use machine learning to model complex biological circuits, achieving speeds of hours instead of years or months. By training a deep neural network on large datasets, they uncover patterns and interactions between variables that were previously impossible to discover.
Researchers at UCSB have enabled identifying a person behind a wall using only WiFi transceivers outside, utilizing received power measurements of a WiFi link. The proposed approach achieves an accuracy of 84% in correctly identifying the person behind the wall, with potential applications in surveillance and security.
A new algorithm developed by MIT student Trevor Henderson seamlessly blends different audio signals in real-time, recreating the portamento effect. The algorithm uses optimal transport to map pitches from one signal to another, producing a smooth transition without volume fading.
Researchers at Queen Mary University of London used AI to analyze 17,000 healthy UK Biobank volunteers' heart MRI images, finding genetic factors account for 22-39% variation in left ventricle size and function. The study identified 14 genome regions associated with heart structure and function.
Duke University engineers used machine learning to design dielectric metamaterials that absorb and emit specific frequencies of terahertz radiation, reducing calculation time from over 2,000 years to just 23 hours. The new designs enable thermophotovoltaic devices that convert waste heat to electricity with higher efficiency.
A study found that cardiovascular disease risk prediction algorithms may underestimate the risk for people with severe mental illness, often defined as schizophrenia, bipolar disorder, and major depression. The algorithm consistently underestimated cardiovascular disease rates in this group, particularly in women.
Researchers at the University of Bern have been awarded prestigious starting grants for their innovative projects on digital 3D models and fluid-rock interactions. The €1.5 million AlgoHex project aims to develop algorithms for automatic hexahedral mesh generation, while the €2 million PROMOTING project seeks to model fluid pathways in...
A three-year grant of 1.2 million rubles will support PhD candidate Angelina Markina's work on efficient algorithms for microstrip antenna design. The project aims to develop mathematical models and prototypes for various frequencies, enabling multi-bandwidth applications including 5G.
Researchers at USC Institute for Creative Technologies found that spontaneous facial expressions are often context-dependent and do not accurately reveal intentions. The study challenges the assumption that facial expressions can be used to predict feelings and actions.
A study by Karolinska Institutet shows that a mobile phone-based treatment adjustment algorithm can improve symptoms and medication adherence in uncontrolled asthma sufferers. The digital tool, AsthmaTuner, analyzes lung function and symptoms to provide personalized feedback and treatment recommendations.
A study of 3.5 million books found that words describing women tend to focus on physical appearance, while those describing men refer to behavior and personal qualities. The analysis highlights the potential impact of biased language on AI systems and employee recommendations.
A new algorithm developed by University of Waterloo researchers increases the efficiency of solar photovoltaic systems and reduces power waste, with potential savings of up to 138.9 kWh/year for small home-use systems. The technique could lead to substantial reductions in emissions from large-scale solar farms.
A team of scientists at EPFL developed an algorithm that can estimate a microscope's resolution from a single image, boosting image quality and enabling optimized imaging conditions. The algorithm has been made available as an open-source plugin, allowing researchers to directly obtain the estimate and optimize their microscopes.
A WPI mathematician is working on a project to develop methods to detect flaws in carbon nanotube materials used in composite rocket fuel tanks. He has developed an algorithm that increases the resolution of density scanning systems nine times, enabling more accurate images of the material's uniformity.
Researchers at the University of Washington have developed an algorithm that can monitor factory and warehouse workers' behavior in real-time, providing risk scores and suggestions for improving ergonomics. The app will alert workers to high-risk actions and provide feedback on how to modify their behaviors.
A team of scientists at Bar-Ilan University has developed a new type of ultrafast artificial intelligence algorithm based on the slow dynamics of brain function. This breakthrough outperforms traditional machine learning algorithms in various fields.
University of Utah chemists developed an algorithm that analyzes previous chemical reaction data to predict hypothetical reactions, narrowing the range of conditions needed for successful synthesis. The model successfully predicted outcomes for various reactions, offering a time-saving solution for pharmaceutical and materials research.
A team of UCI researchers developed a deep reinforcement learning algorithm called DeepCubeA, which can solve the Rubik's Cube in under 20 moves, outperforming human solvers. The algorithm works on other combinatorial games and demonstrates symbolic, mathematical, and abstract thinking capabilities.
A Carnegie Mellon AI program called Pluribus defeated leading poker professionals in six-player no-limit Texas hold'em, achieving superhuman performance. It developed unique strategies such as frequent donk bets and mixed play, which human pros struggled to execute.
Researchers use Word2vec to analyze relationships between words in scientific literature, predicting discoveries of new thermoelectric materials years in advance. The algorithm learns concepts like the periodic table and crystal structure of metals without human guidance.
A new AI tool called EternaBrain uses a neural network approach to predict the choices of top players in an internet-based videogame. The researchers discovered that EternaBrain outperforms random guessing and performs similarly or better than previously developed algorithms.
Cornell researchers develop algorithm to visualize models of the universe, unlocking secrets of cosmology and dark matter. The algorithm uses intensive principal component analysis to extract patterns from large datasets, providing new insights into the nature of our universe.
A new AI-powered tool can identify triatomine bugs, also known as 'kissing bugs', that spread the incurable disease Chagas with high accuracy, allowing public health officials to target prevention efforts. The algorithm achieved 95.8% and 98.9% correct identification rates for Mexican and Brazilian species, respectively.
Researchers at MIT develop a new algorithm that can accurately pick out an object, such as a small animal, in a dense cloud of dots within seconds. The technique prunes away outliers quickly, even for increasingly dense clouds, making it suitable for applications like driverless cars and robotic assistants.
Researchers improve consensus in Proof of Stake protocols by introducing weighted voting schemes, ensuring fairness and security. The new approach scales votes according to validators' profiles without interfering with existing PoS execution.
A Massachusetts Institute of Technology-developed system automatically generates optimized cryptography code for Google Chrome browsers and web applications, matching performance of handwritten code but faster. The system, known as Fiat Cryptography, is being widely used by tech firms and has populated Google's BoringSSL library.
A new AI tool developed by researchers at Stanford University improves clinicians' ability to correctly identify brain aneurysms by highlighting areas of interest on scans. The HeadXNet algorithm reduces the 'miss' rate and increases consensus among clinicians, with promising results but further investigation needed.
A new microgrid system designed by American and Chinese researchers promises improved stability, safety, and resilience in delivering energy. The system aligns several energy sources in parallel and uses a decentralized control algorithm to overcome the burdens of system overload and shutdown.
Researchers at RIT are developing a computational framework to reconstruct and visualize myocardial active stress, enabling direct appraisal of cardiac function. This will help detect and localize regions with diminished contractile activity, improving diagnosis and treatment of cardiac disease.
Researchers develop a new algorithm using over 230,000 patient records to predict mortality risk and survival chances. The algorithm uses disease history from the past 23 years to make more accurate predictions than existing methods.
Researchers at Pitt and CMU aim to develop AI-powered robotic suits that can treat and stabilize soldiers in the field, extending the 'golden hour' for medical evacuations. The project uses real-world data and medical algorithms to guide critical care interventions.
USC researchers Bhaskar Krishnamachari and Aditya Asgaonkar propose a blockchain-based solution to reduce costs and increase efficiency in online transactions. Their system allows buyers and sellers to interact directly without third-party mediators, using smart contracts to verify honesty and disincentivize malicious behavior.
A new brain-controlled hearing aid technology developed by Columbia engineers can identify and amplify the correct speaker in a crowded environment. The device uses artificial intelligence to monitor wearers' brain waves and boost the voice they want to focus on, solving the 'cocktail party problem' that modern hearing aids struggle with.
A new algorithm called DeeProtein uses sensitivity analysis to unravel the secret of its predictions, providing valuable insights into protein functions. This technique enables researchers to identify critical regions in proteins that tolerate changes well or poorly, paving the way for targeted modifications.
A machine learning algorithm can detect signs of anxiety and depression in young children's speech, potentially providing a fast and easy way to diagnose conditions that are difficult to spot. The algorithm is highly successful at diagnosing children with an internalizing disorder with 80% accuracy.
Researchers developed an algorithm called Scanorama that merges over 20 diverse human cell datasets into a single, diverse source of data. The algorithm uses a modified computer-vision technique to find matching cells across datasets and preserves unique cell types.
A team of researchers has developed a mathematical model to calculate the cost - time and energy - to complete a task based on the number of drones and recharging stations available. The model considers the energy required for each drone to complete its portion of the task and fly to a charging station as needed.
Researchers have developed an energy renormalization algorithm to predict glass' mechanical behavior at varying temperatures. This approach enables the design of dynamic materials with optimal properties, scaling molecular simulations up by roughly a thousand times.
Researchers at Rutgers University have developed a low-cost, high-speed algorithm to test chemical toxicity without animals. The algorithm uses massive amounts of data from PubChem and achieves a 62-100% success rate in predicting oral toxicity for several groups of chemicals.
Researchers propose a unified approach to optimize radar polarimetry data for precise weather forecasting. This combines observation-based retrievals with model-based analysis to improve quantitative precipitation estimation, warnings, and forecasts.
A new AI approach analyzes time-lapse images of embryos to predict pregnancy potential, improving IVF success rates and minimizing multiple pregnancies. Researchers used a large dataset of human embryos to train an algorithm that can classify embryo quality with high accuracy.
A new mathematical method developed by USC researchers can identify anomalies in self-driving cars' perception algorithms before they hit the road, improving safety. The method uses 'sanity conditions' to test machine learning tools and can be used to pinpoint specific problems and retrain the algorithms for faster error detection.
A team of healthcare data scientists and doctors developed an AI system using machine learning algorithms to predict the risk of early death due to chronic disease. The system performed better than current standard approaches and showed promise in improving preventative healthcare.
Researchers used an algorithm to model the decision-making of C. elegans in response to a sensory stimulus, achieving predictions that matched experimental results. The Sir Isaac platform demonstrated improved accuracy compared to prior models, offering insights into the potential of artificial intelligence in scientific discovery.
The Blue Brain Project has developed an algorithm to objectively classify the shapes of neurons in the brain, enabling the creation of a standardized taxonomy of all brain cells. This breakthrough resolves a century-old neuroscience problem and provides a reliable comparative method for researchers.
Researchers at MIT have developed a neural architecture search algorithm that can directly learn specialized convolutional neural networks for target hardware platforms in only 200 GPU hours. The algorithm uses 'path-level' binarization and pruning to reduce memory consumption and improve efficiency.
Researchers at University of California, Davis and Maynooth University created programmable DNA molecules that can self-assemble into patterns by running their own program. They designed and ran 21 algorithms, demonstrating the potential of the system for sophisticated molecular engineering.
Researchers design DNA molecules that can carry out reprogrammable computations, executing six-bit algorithms for simple tasks. The system demonstrates versatility and potential for teaching about molecular processes encoding information.
Researchers from Mitsubishi Electric Research Laboratories developed an improved algorithm to track motor performance and speed estimation without sensors. The proposed algorithm uses state variables to estimate rotor speed, addressing limitations in existing approaches.
A new study published in Ophthalmology shows that AI can assist physicians in accurately diagnosing diabetic retinopathy, a potentially blinding eye disease. The researchers developed two types of assistance to help physicians read the algorithm's predictions, which improved diagnostic accuracy and confidence.
Researchers at Garvan Institute of Medical Research have developed a computational method to reduce the amount of memory necessary for genome alignment, allowing for real-time analysis on smartphones. This breakthrough enables remote disease identification and point-of-care microbial infections.
Researchers have created a novel way to define individual protein associations using the topological scoring (TopS) algorithm. This approach helps identify proteins that come together in biological processes, allowing for better understanding of how proteins perform functions and interact with each other.
Omri Weinstein, assistant professor at Columbia University, has won a National Science Foundation CAREER Award to explore data structures and information retrieval. His research aims to develop new mathematical tools for proving lower bounds on data structure efficiency and scalable storage applications.
Researchers from the Center for Genomic Regulation developed a method to predict and classify these tiny proteins using bioinformatics tools, discovering they account for 16% of bacterial genomes. The small proteins play a crucial role in antimicrobial responses, microbiota balance, and may be overlooked in complex organisms.
NIST researchers demonstrate deep learning algorithms outperform traditional methods for detecting offshore radars, improving spectrum sharing. The new approach provides occupancy statistics for the 3.5 GHz band, enabling commercial users to determine when to yield to naval operations.
Rice University researchers have developed a method to capture 4D data using 2D microscopes, enabling scientists to visualize molecules' locations and movements in living cells. The technique uses custom phase masks to manipulate light and separate spatial and temporal information.
A University of Texas at Arlington researcher is leading a high-powered team in the Advanced Research Projects Agency-Energy (ARPA-E) Grid Optimization Competition. The team has secured $250,000 in funding and aims to develop software solutions for efficient energy delivery through the nation's power grid.
A machine learning algorithm has been developed to speed up the process of discovering new medicines, identifying four new molecules that activate a protein relevant to symptoms of Alzheimer's disease and schizophrenia. The algorithm is twice as efficient as industry standards and can analyze vast amounts of chemical data.
A new study finds that women's brains are metabolically three years younger than men's of the same age, which may contribute to their greater mental sharpness in later years. The researchers used PET scans and machine-learning algorithms to measure brain metabolism and calculate each person's brain age.