Articles tagged with Algorithms
A new algorithm developed by Tufts University researchers enables the self-localization of IoT devices on 5G networks, eliminating the need for centralized anchors and improving location-based services. This scalable solution can meet the demands of a projected 50 billion connected products by 2020.
Researchers at MIT have developed a fleet of autonomous boats that can ferry goods and people, reducing street traffic. The boats can also be printed using 3D technology and equipped with sensors to monitor water quality, providing insights into urban health.
A new lane-change algorithm developed by MIT allows autonomous vehicles to model an entire range of driving styles, from conservative to aggressive, with safety guarantees. The system computes buffer zones on the fly and ensures collision avoidance within a given time frame.
Researchers developed a computational way of analyzing X-ray images of lungs to assess Chronic Obstructive Pulmonary Disease (COPD) and other lung diseases. The new method uses topology, combining CT scans, high-performance computing, and algorithms to identify lung function characteristics.
HRL Laboratories joins DARPA's Assured Autonomy program with the Expressive Assurance Case Toolkit (ExACT) to ensure autonomous vehicle systems perform as programmed without unsafe behavior. The tool kit mathematically verifies algorithms lead to safe and reliable system behavior, considering physics and dynamics of the system.
A study by Columbia University researchers found that social networking site Instagram's recommendation algorithms amplify homophily, making women less visible and reducing their popularity. The algorithms, which turn loose on a network with homophily, effectively make women less visible.
A new open-access software called PyFRAP has been developed to accurately analyze molecular diffusion in living cells. The program takes into account three-dimensional structures and provides reliable results, especially under complicated conditions.
Daniel Sheldon's NSF grant aims to develop efficient algorithms for complex models, enabling scientists to make predictions and test hypotheses from large datasets. He plans to address challenges in modeling animal populations, causal reasoning about citizen scientist data, and optimizing decision-making with multiple objectives.
Researchers investigate how algorithms create personalized content and 'filter bubbles' that can foster biases and inequality. The MyBubble project aims to provide transparency into online advertising and profiling.
Researchers at University at Buffalo School of Management developed a new algorithm to establish pricing and allocate resources for cloud computing. The study provides practical formulas that companies can use to determine necessary resources and set charges for their customers.
A new method developed by researchers at USC Viterbi School of Engineering uses machine learning to measure key risk factor for cardiovascular diseases and arterial stiffness using just a smartphone. The method was validated with existing tonometry data and showed high correlation with actual tonometry measurements.
Researchers at UC Berkeley have developed a method using deep reinforcement learning to create realistic computer animations. The system, dubbed DeepMimic, can simulate various skills, including acrobatic feats like break dancing and martial arts, and respond naturally to changes in the environment.
A new 'smart' algorithm is proposed to manage household energy usage, taking into account varying power sources like windmills and solar panels. The goal is to minimize costs while meeting energy demand, with the potential for future optimization through machine learning.
A study using machine-learning algorithms and MRI analysis predicts the severity of neurologic impairment after spinal cord injury. The results show variable accuracy in classifying patients and identifying potential prognostic texture features.
Scientists from Spain have created a new technique using magnetic resonance imaging (MRI) to evaluate the properties of Iberian hams and whole loins. This non-invasive method allows for the prediction of quality characteristics such as fat content, moisture levels, color, and sensory attributes without destroying the meat products.
Researchers created an algorithm that sorts intervention program participants into smaller groups based on social connections and prior history of substance abuse. The AI tool performs significantly better than control strategies in forming groups, suggesting that evenly distributing regular users can decrease success rates.
A new study by UTSA researchers presents a method to stop cyber attacks on GPS-enabled devices, including electrical grids and driverless cars. The algorithm can recognize false GPS signals and counter an attack in real-time, making it crucial for securing GPS signals used in various technologies.
Researchers at UPV/EHU developed an algorithm to calculate chest compression depth and frequency based on chest acceleration, validating its accuracy in real-time monitoring of CPR quality. A new device is being marketed to assist CPR, aiming to improve patient survival rates.
The eGVHD App is a user-friendly tool that supports accurate GvHD diagnosis and scoring after HCT. It provides guidelines, definitions, and pictures to help clinicians evaluate all potentially affected organ systems.
Researchers at U.S. Army Research Laboratory have developed a way to factor large composite integers using brain-inspired computer architectures, moving away from traditional computing methods. This breakthrough could break many modern-day internet security protocols, including public key encryption.
The use of artificial intelligence in medicine raises important ethical questions, including bias in algorithms and the potential for data to replace human clinical experience. Researchers call for careful examination of these risks and for physicians to critically assess machine-learning tools and their limitations.
The study, led by George Mohler, found no statistically significant difference in arrest rates by ethnic group between predictive policing and standard patrol practices. The researchers also discovered that arrests were higher in algorithmically-selected areas but remained unchanged when adjusted for crime rates.
A field-data study in Los Angeles found that predictive policing did not result in biased arrests, with arrest rates by ethnic group showing no statistically significant difference. However, the researchers noted that higher crime rates and proportionally higher arrest rates were observed in algorithmically selected areas.
A team of researchers has developed an algorithm that uses GPS tracking and data analytics to optimize the harvesting of hand-picked crops. By analyzing the movement of each worker in real-time, the algorithm can predict the amount of completed boxes and improve overall efficiency.
Researchers at Lawrence Berkeley National Laboratory developed HipMCL, an algorithm that can cluster large biological networks containing millions of nodes and edges. The new method allows biologists to make sense of big science data using massively parallel supercomputers.
A new study has developed an automated approach to select and seed teams for the NCAA March Madness basketball tournaments, providing an objective and unbiased starting point. The algorithm achieved high accuracy rates, matching 24 of 38 top teams in seeding and 89.6% being within two seeds of committee rankings.
A new study presents a machine learning algorithm that accurately predicts heart attack diagnoses with a 94% accuracy rate, outperforming standard clinical models. The algorithm uses large collections of patient data to identify patterns and improve routine processes in healthcare delivery.
A USC-developed algorithm can better locate and treat people living with undiagnosed infectious diseases, such as TB and gonorrhea. The algorithm optimizes disease spread models using behavioral, demographic, and epidemic trends to capture underlying population dynamics.
A team from Lehigh University's Industrial and Systems Engineering department participated in a workshop on big data optimization algorithms, theory, applications and systems. Researchers presented their work on Interior Point Methods, machine learning methodologies, and other topics relevant to big data analytics.
NanoMap enables drones to fly through cluttered spaces like cities at high speeds using a simple yet effective approach that considers the drone's position in the world over time. The system reduces crash rates from over 90% to just 2%, making it suitable for applications such as search-and-rescue, defense, and package delivery.
Researchers at USC Center for Artificial Intelligence in Society developed an AI algorithm, SPOT, to detect poachers in near real-time using infrared images. The algorithm can distinguish between humans and animals in just over three-tenths of a second, reducing the time spent on monitoring wildlife.
Researchers developed DroNet, an algorithm that enables drones to safely navigate through urban environments using a normal camera and powerful artificial intelligence. The drones learned to respect safety rules and adapt to new environments, paving the way for fully autonomous drones in various applications.
A team of computer scientists developed an algorithm called VarQuest to rapidly search massive databases for new antibiotic variants, discovering over a thousand new variants.
The Northwestern University team developed highly efficient metadevices at millimeter-wave frequencies using inverse design principles and 3D printing. This approach starts with a function and asks what structure is needed to achieve the desired result, producing unexpected outcomes like broad bandwidth functionality.
A new algorithm developed by Stanford University's Immigration Policy Lab can optimize refugee resettlement, leading to improved employment and integration rates. The algorithm takes into account individual characteristics and local conditions to match refugees with the most suitable locations.
A new algorithm developed by Stanford researchers can help resettle refugees more effectively, improving their employment success and overall integration. The algorithm analyzes historical data on refugee resettlement and assigns placements that project a 40-70% increase in employment rates compared to actual outcomes.
A recent study by Brigham Young University researchers found that machines trained with an algorithm called S# outperformed humans in finding compromises that benefit both parties. The machines were able to learn from 'cheap talk' phrases and maintain cooperation, often more effectively than humans.
A Dartmouth College study found that non-experts who responded to an online survey performed equally as well as the Correctional Offender Management Profiling for Alternative Sanctions (COMPAS) software system in predicting repeat criminal behavior. The study also demonstrated that only two variables - a defendant's age and number of p...
Researchers used digital pathology slide images to train computer algorithms that could detect the spread of breast cancer to lymph nodes with high accuracy. The study found that AI algorithms outperformed human pathologists in detecting cancer spread, suggesting potential improvements in diagnosis and treatment outcomes.
A new recommendation algorithm uses cosine similarity to predict customer preferences in high-dimensional spaces, leveraging the angle between user rating lines. The approach balances signal gain with noise introduction, resulting in more accurate predictions and better performance, particularly when ratings data is sparse.
A new algorithm, developed by a team at the University of Maryland, uses artificial neural networks to address multiple flaws in a single image. The algorithm can be trained on high-quality images and then applied to any image with imperfections.
A new algorithm can speed up protein-folding simulations, allowing researchers to model phenomena that were previously out of reach. This technique can help scientists better understand and treat diseases like Alzheimer's, which is associated with amyloid-beta protein fragments forming hard plaques that disrupt neurons.
The Meta Planning Engine system detects changes in demand and uses multiple algorithms to find the best solution for traffic management. It provides flexibility and adaptability, allowing the software to choose the optimal algorithm for each situation.
Researchers at Lobachevsky University have developed a mathematical model to explain cognitive dissonance through the lens of information images. The theory posits that there is a limited space filled with interacting information images, governed by laws that determine real and virtual activities.
A Worcester Polytechnic Institute researcher has developed algorithms to identify fake likes, followers, and posts on various platforms with high accuracy. The goal is to prevent crowdturfing, which can spread misinformation and undermine online credibility.
Researchers developed Fastron, a machine learning-based collision detection algorithm that runs up to 8 times faster than existing methods. The algorithm uses a minimalistic approach to classify collisions versus non-collisions in dynamic environments.
Researchers at Salk Institute have found that fruit fly brains use an efficient method to perform similarity searches, expanding the dimension of odor information to improve detection. This approach could inform computer algorithms and enhance their ability to find similarities quickly.
A researcher at the University of Kansas is working on developing a foundational mathematical framework to account for unstructured uncertainty in complex dynamic systems. The goal is to improve predictive accuracy in fields such as engineering, nature, and global climate change.
A new computer model, Recursive Cortical Network (RCN), has been developed that can solve CAPTCHA and identify handwritten digits with comparable accuracy to state-of-the-art deep learning approaches but using significantly less training data.
A team of scientists and amateur astronomers have detected six exocomets, the smallest objects yet found outside our solar system, using transit photometry. The detection marks the first time an object as small as a comet has been identified using this technique.
Researchers from University of Leicester and KU Leuven found that information flow accounts for emergence of small-world networks in complex systems. These structures arise spontaneously in neuronal and social networks, and are characterized by short-cuts and hierarchical organization.
Computer scientist Jelani Nelson finds Johnson-Lindenstrauss lemma best approach to reduce data dimensionality. The theorem helps speed up algorithms across various fields.
Researchers use angular correlations of scattered X-rays to determine the 3D structure of biological objects, overcoming limitations of traditional X-ray methods. The breakthrough enables scientists to study molecular structure and dynamics previously impossible to observe.
Scientists discovered a network of neurons in bees' brains that integrates direction and distance changes, enabling them to return home. A detailed computer model was developed based on this research, which could lead to new algorithms for autonomous robots.
Researchers have discovered a potential method to predict nearby strong earthquakes by analyzing deep tremors. The study, published in the Journal of Geophysical Research: Solid Earth, found that changes in deep tremor patterns can signal an impending earthquake.
Chinese scientists develop a new algorithm that leverages network structure characteristics to improve link prediction accuracy and robustness. Their experimental testing in various real-world networks yields better results than existing methods, leading to the creation of a novel method for predicting missing links.
Four primary methods are used by primary care practices to risk stratify patient populations: practice-developed algorithm, AAFP clinical algorithm, payer claims/electronic health record, and clinical intuition. Practices that developed their own algorithm identify more high-risk patients than those using other methods.
Researchers developed an algorithm to generate contiguous district shapes, balancing population and political affiliation. The tool enables state legislatures to explore options in a transparent and efficient manner, promoting fairness and reducing partisan biases.
The system uses cameras and AI to detect hand movements deviating from normal driving behavior, grading them for possible safety threats. Researchers hope the technology will reduce traffic accidents caused by distracted drivers, a major contributor to global crashes.
A computer algorithm developed by ITMO University's programmers analyzes Instagram photos taken in Saint Petersburg to identify popular locations favored by residents. The results provide insider information for tourists, offering a more authentic experience.