Articles tagged with Algorithms
An expert system can help diagnose schizophrenia's early onset symptoms, allowing for timely referrals and earlier treatment. The system uses AI to process reported symptoms and identify potential cases.
Researchers at Northeastern University have cracked the code of Uber's surge pricing algorithm, finding that prices can vary wildly based on location and demand. The team has developed a detailed map of surge areas, revealing that some areas offer significant discounts by walking or waiting five minutes.
Researchers have developed a new cutting-plane algorithm that improves upon existing methods, offering huge efficiency gains for specific optimization problems. The algorithm has been applied to various types of problems, including submodular minimization and semidefinite programming, resulting in significant speedups.
Researchers have presented new unconditional hardness results for dynamic and online problems, including matrix-vector multiplication and Patrascu's Multiphase Problem. These findings prove the existence of problems that are provably hard to solve, offering significant advancements in computer science.
Researchers have developed an algorithm that can analyze a person's intention despite interruptions or disturbances, allowing for more accurate control of robots and prosthetics. This technology could be used to guide cars on icy roads or assist stroke patients with mobility issues.
New algorithms have been developed using the natural process of Japanese tree frogs singing out of sync to solve computational problems in graphs and social networks. These algorithms, known as FrogCOL and FrogMIS, can identify independent sets of nodes and detect structural patterns in complex networks.
Researchers at MIT develop a new way of doing machine learning that enables semantically related concepts to reinforce each other. In experiments, their system outperformed conventional training methods in predicting Flickr tags and even when considering semantically similar tags.
The new algorithm enables filmmakers to determine the optimal number and location of cameras for capturing a scene, amassing data that can be used during post-production. It allows for virtual relighting, generating new viewpoints, and digitally recovering detailed 3D shapes and layouts.
The DOE is investing $10.5 million in CAMERA to develop computational mathematics and algorithms for experimental facilities research. This will help scientists extract useful information from experiments, tackling challenges such as data analysis and image processing.
A new algorithm using a highly sensitive troponin I assay can reduce mortality and cut triage times to one hour for patients with suspected acute myocardial infarction. The algorithm allows for safe discharge or rapid treatment initiation, enabling better resource utilization.
Researchers have created an algorithm that can efficiently search for the laws underlying complex dynamical systems, including biological processes. The algorithm, dubbed 'Sir Issac', was tested on artificial and real-world systems, demonstrating its ability to infer universal gravitational force and model chemical transformations.
A new study from NYU uses 7.7 million 311 service requests to track where and when neighborhood conflict arises, finding it more likely in areas with 'fuzzy' boundaries between homogenous communities. Conflict is less common in areas without clear boundaries or with clearly defined community groups.
A team of computer scientists from University of Utah discovered a way to find bias in algorithms used for hiring decisions and loan approvals. They developed a technique to determine if such software programs discriminate unintentionally, revealing potential issues with disparate impact. The researchers also found a method to fix thes...
Researchers developed an algorithm to analyze and critique digital photographs' composition, providing feedback on perceived aesthetics. The system relies on psychological theories of human vision and opinions of people, helping photographers improve their work.
The researchers developed a new algorithm using geometric matching and machine learning to analyze style similarity in 3D objects. They achieved an agreement rate of 90% with human evaluations, making it a valuable tool for computer graphics applications.
A new mathematical theory and algorithm, Multi-tiered iterative phasing (M-TIP), solves the reconstruction problem for fluctuation X-ray scattering data. This approach enables quick determination of general structure in minutes on a desktop computer, unlocking new advances in biophysics.
Researchers developed a 'reusable holdout' tool to identify meaningful correlations in adaptive data analysis. The algorithm, tested on two rigged datasets, correctly identified true statistical significance and eliminated false discoveries.
Researchers discovered that redeposition acts as a stabilising factor in the creation of hexagonally arranged dot patterns. The study reveals that eroded particles predominantly redeposit in valleys, not at hilltops, contributing to stable nanostructure formations.
A team led by Stanford electrical engineer Krishna Shenoy has developed a technique to make brain-controlled prostheses more precise, enabling people with spinal cord injuries to tap out commands with greater accuracy. The new approach continuously corrects brain readings to give users a more natural way to interact with devices.
Jennifer Tour Chayes of Microsoft receives the highest honor from SIAM, awarded for her groundbreaking work on phase transitions in mathematics and computing. She will deliver a prize lecture on massive networks on August 12.
Researchers developed an algorithm inspired by brain pruning, which eliminates unneeded connections in early childhood. The neuroscience-based approach created more efficient and robust networks than traditional engineering methods.
The book explores how hierarchical organization, wavelet transformation, subspace trees, and deep learning can overcome the curse of dimensionality to develop efficient big multimedia databases. It introduces essential statistical supervised machine learning algorithms for information retrieval.
A new learning method for energy optimization of plug-in hybrid electric buses uses stochastic Markov models to minimize fuel and electric consumption costs. The approach avoids the curse of dimensionality problem by approximating the cost function without discretization, reducing computational complexity.
Researchers developed a dropout-prediction model that uses data from one course offering to predict stopout in the next. The model achieved fairly accurate predictions and showed promise, particularly when incorporating additional variables like weekend study habits. Ongoing work aims to refine the model for improved accuracy.
A study using NASA satellite data found that collecting rainwater for vegetable irrigation in India can reduce water bills by up to 4,522 rupees per year. This can provide a profit of between 1,548 and 3,261 rupees per year, boosting cost savings and increasing quality of life.
Researchers at Hiroshima University found a correlation between telomere G-tail length and endothelial dysfunction, as well as stroke and dementia, in patients with chronic cerebrovascular disease. The study suggests that telomere G-tail length may be a useful biomarker for predicting cardiovascular risk.
Researchers at MIT report that a widely held assumption about computational complexity is correct, making it unlikely that a more efficient algorithm exists. The current edit distance algorithm, in use for over 40 years, has been deemed the best possible solution for comparing genomes or strings of symbols.
A new algorithm developed by Microsoft researchers has improved route planning capabilities for Bing Maps, providing faster processing and more accurate information on lengthy trips. The Customizable Route Planning (CRP) engine also offers more reliable alternate routes and better estimates of time needed for turns and traffic snarls.
A new MIT algorithm significantly reduces robot teams' planning time by deferring difficult grasp decisions until easier ones are made. The algorithm produces a workable plan that may not be perfectly efficient but saves time in planning, often offsetting added execution time.
Researchers at UC Berkeley developed algorithms that enable robots to learn motor tasks through trial and error, mimicking human learning. The 'deep learning' technique allows robots to adapt to changing environments without pre-programmed details.
A new study by Douglas Mental Health University Institute researchers suggests that the shape of the hippocampus, not its size, is linked to improved memory function. The research, published in Human Brain Mapping, found that individuals with a broader hippocampus tend to perform better on memory tests than those with a smaller one.
Behnaz Ghoraani's novel algorithm allows improved and patient-specific localization of electrical disturbance sites, enabling clinicians to better target the problem. The technology helps doctors 'see' what's going on in the atrium by analyzing patterns of electrical signals, guiding them to where they need to sense.
A neural network model correctly predicted 96% of US bank bankruptcies between 2002 and 2013. The model identified construction loans and rapid growth as key risk factors for bank failure.
Computer scientists developed a new method combining computer vision algorithms and brain-computer interface to detect mines in sonar images, outperforming existing methods. The system uses classifiers to capture changes in pixel intensity and improves accuracy by detecting 99.5% of true positives and reducing false positives.
A team of US computer scientists are working on an analyzer that can hunt down algorithmic attacks, a new form of vulnerability that is nearly impossible to find with today's technology. The software will perform a mathematical simulation of the software to predict what will happen in the event of an attack.
Scientists have developed a new method to monitor Europe's nature protection areas by analyzing reflected light signals from laser pulses. The technique allows for the identification of different types of vegetation, including weeds and vehicle tracks, and provides detailed 3D maps of the landscape.
Researchers at MIT's CSAIL have developed a new algorithm, DIODE, to detect integer-overflow bugs in programs. The system works by feeding a single sample input and recording each operation performed on it, allowing it to identify potential security vulnerabilities.
A new computational tool identifies 800 ways people are at increased risk for post-traumatic stress disorder (PTSD), allowing for personalized prediction guides. The algorithm, developed by NYU Langone Medical Center researchers, can predict PTSD risk less than two weeks after a traumatic event.
Researchers at the University of Michigan have developed an open-source digital assistant called Sirius, which uses advanced algorithms to interpret voice commands and answer questions in spoken words. The system has the potential to revolutionize the development of wearable devices and datacenter infrastructure.
Research reveals that cities in Spain have a 15-year memory, while those in the US have a 25-year memory, influenced by historical events. The study also shows that city growth is linked to neighboring cities and their surrounding areas.
Researchers at MIT developed a system that cleverly distributes data around multicore chips' memory banks, improving execution times by 18 percent on average while increasing energy efficiency. In simulations involving a 64-core chip, the system increased computational speeds by 46 percent while reducing power consumption by 36 percent.
A team of researchers has created a high-resolution map of the Ophiuchus molecular cloud, identifying 75 dense cores and finding that their masses follow a log-normal distribution. This suggests that turbulence plays a significant role in shaping the evolution of the cloud.
A new way of implementing priority queues allows for performance improvement even with the addition of new cores. The algorithm uses a skip list data structure to make moving through the hierarchy more efficient, reducing collisions and achieving better scalability.
A new algorithm, TopicMapping, has been developed to improve the accuracy and reproducibility of big data text analysis. By using a network approach, the algorithm separates unstructured text into topics with high accuracy and reproducibility. The results show that existing algorithms, such as LDA, are not reliable for complex datasets.
Researchers at MIT developed an algorithm to optimize optimization algorithms, guaranteeing the best possible solution for complex engineering problems. By using a Gaussian smoothing technique, they generate a sequence of simpler problems that progressively add complexity, ensuring convergence to a global minimum.
Researchers at MIT develop an algorithm that allows household robots to recognize four times as many objects as traditional methods, reducing misidentifications. The new approach uses a representative sampling technique, sampling hypotheses across successive images, and evaluating them more efficiently.
A new algorithm allows for the efficient decomposition of 3D objects into pyramidal parts, minimizing material waste and saving print time. This breakthrough has significant implications for 3D printing, molding, and casting, enabling the creation of complex shapes with minimal support material.
A new system developed at MIT enables pattern-recognition systems to distill what they learn into simple examples, which humans can use to make better decisions. In experiments, human subjects using the system outperformed those using a similar algorithm by over 20 percent.
Researchers warn that social media datasets may be misleading due to built-in population biases and algorithmic changes. Scientists need to correct or acknowledge these biases to ensure accurate results.
Candès will be awarded the 2015 AMS-SIAM George David Birkhoff Prize for his groundbreaking work on compressed sensing, a mathematical technique that has revolutionized signal processing and medical imaging. His research has led to dramatic advances in data collection and analysis, with applications in MRI machines and digital cameras.
Research at Princeton University and New York University reveals multiple pathways for phase transitions, contradicting previous theories. The study suggests that phase changes involve point defects and disordered clusters, leading to a more nuanced understanding of matter's behavior.
Researchers at MIT have developed a new system, Chisel, that enables programmers to prioritize energy savings over computational accuracy in certain applications. The system uses a tool to help programmers evaluate the acceptable level of error and automatically assigns instructions to unreliable hardware components.
A research team at UMass Amherst aims to understand the complex interactions between content providers, ISPs, and CDNs, with the goal of improving online video streaming quality. The researchers plan to apply interdisciplinary techniques from control theory and network algorithmics to develop new architectures for better coordination.
Researchers found that people respond to surveys by selecting answer options with similar language, even when the similarities are subtle. The study highlights the importance of considering language patterns in survey design and analysis, potentially revealing unique attitudes among respondents.
The Gauss von Mises (GVM) distribution offers improved predictive capabilities for tracking infrequently-observed space objects. This new approach allows for more accurate prediction of satellite and debris locations, enabling better resource allocation and detection of potential collisions.
Researchers equipped a robot with a novel tactile sensor, allowing it to grasp and manipulate objects in unprecedented ways. The sensor uses optics and computer-vision algorithms to infer the three-dimensional structure of surfaces, giving the robot feedback in real-time.
MIT researchers developed an algorithm to gauge space debris rotation using visual information, achieving high accuracy in most measures. The algorithm's probabilistic approach and Gaussian distribution modeling enable efficient collection of space trash, including thousands of broken satellites.
Researchers at MIT developed an analytic framework that enables the analysis of GPS data on 150 million cab rides in New York City. They found that if passengers were willing to tolerate delays of five minutes or less, almost 95% of trips could have been shared.
A novel decision support system combines computer vision algorithms with visualization of relevant areas to aid in malaria diagnosis. The method was shown to be accurate comparable to World Health Organization quality criteria, with over 90% of infected samples correctly diagnosed.
A new algorithm developed by Hua-Wei Shen helps accurately allocate credit to researchers based on their contributions. The algorithm, which has been tested on 63 Nobel prize-winning papers, found that authors deserving of the most credit corresponded to the Nobel laureate in 81% of cases.