Articles tagged with Data Points
Researchers developed a neural network that can adapt to new data inputs, continuously learning from changing time series data streams. This 'liquid' network could boost the development of emerging technologies like self-driving cars and medical diagnostics.
A new study found that anonymous cell phone data can quantify behavioral changes for flu-like illnesses, while protecting user anonymity. The research linked passively-collected cell phone metadata with actual public health data, providing a potential tool to monitor and control infectious disease outbreaks.
UT Arlington computer scientists develop a deep learning method to generate synthetic objects for robot training, overcoming the need for manual capture of images from human-centric perspectives. The technique uses generative adversarial networks (GANs) to create photorealistic full scenes and dense colored point clouds with fine details.
Researchers from Lund University criticized a model used to evaluate lockdowns, stating it had fundamental shortcomings and couldn't be used to draw conclusions. The Imperial College London model attributed almost all reduction in transmission during spring to lockdowns, with Sweden being an exception that offered a different explanation.
Researchers have developed a machine learning model that predicts the hardness of new materials based on their chemical composition, finding over 10 promising borocarbide phases. The model's accuracy surpasses classical methods, enabling scientists to more efficiently discover suitable compounds for various applications.
A new deep-learning algorithm called TadGAN has been developed to detect anomalies in time series data, outperforming traditional methods. The algorithm combines the strengths of generative adversarial networks and autoencoders to strike a balance between vigilance and false positives.
Researchers developed an airborne approach to map live coral distribution in the Hawaiian Islands, revealing a negative correlation between nearshore development and live coral cover. The study suggests that this cost-effective method could provide high-resolution monitoring of coral health at a low cost.
Researchers have developed a new computational method called Wasserstein-induced flux (WIF) to validate the accuracy of images in scientific research. WIF provides statistical confidence that specific data points in an image are accurate, enabling scientists to evaluate model performance and improve their models.
A FAIR Health initiative found that cost information enhances the value of shared decision making in palliative care scenarios, empowering patients to make critical healthcare decisions. The program's success highlights the utility of cost data in guiding informed discussions with clinicians.
Researchers review factors contributing to disconnect between created and utilized marketing data, highlighting overlooked sources for growth opportunities. The study explores the potential of biometric and social network data to support customer acquisition, relationship development, and retention.
A new mattress flammability standard has been shown to prevent 65 deaths annually from bed fires. The standard, established by the National Institute of Standards and Technology (NIST), requires mattresses to maintain a heat release rate under 200 kW after being ignited.
Researchers at the University of Cambridge developed vLUME, a virtual reality software that allows for the visualization and analysis of complex biological data. This enables faster testing and analysis, leading to new discoveries and insights into fundamental problems in biology.
Researchers used data from fishing technology company ReelSonar to model angler movements and predict where aquatic invasives may spread. The study found that anglers who fish frequently in the same area are more likely to transport invasive species.
A new model by Vicky Chuqiao Yang accurately describes why politicians become more polarized despite voters staying moderate. The model, guided by the concept of 'satisficing', assumes people settle for a candidate who is 'good enough' rather than always voting for the best one.
A team of researchers at Argonne National Laboratory has discovered a possible breakthrough solution for controlling defects in 3D printing by using temperature data to predict the formation of subsurface defects. This could eliminate costly inspections and enable mass production of complex metallic components.
The National Academies has published a guide to help public officials interpret COVID-19 data, considering factors such as data representation, biases, and uncertainties. The guide aims to provide decision-makers with a comprehensive understanding of the strengths and weaknesses of various data sources.
Researchers identified distinct immune profiles, 'Immunotypes,' in 125 hospitalized COVID-19 patients. These signatures correlated with disease severity and provided insights into the human immune response.
A new method developed at Carnegie Mellon University has improved the accuracy of tracking systems for autonomous vehicles. By training on larger datasets and detecting its own errors, the system can learn to do the right thing without being told what it is.
Researchers at Carnegie Mellon University developed a perception system that considers visibility to improve object detection in self-driving cars. The method outperformed previous techniques, with improvements of up to 18.4% for buses and 16.7% for trailers.
Researchers have developed a new imaging analysis pipeline called LaMDA, which enables machine learning and complex analyses on high-dimensional microscopy data. This can reveal and predict T-cellular states, providing valuable insights for drug testing, vaccine development, and expanding knowledge of T-cell biology.
A new study found that online symptom checkers are only accurate about a third of the time, and 'Dr. Google' is often incorrect in its diagnoses. The research also found that triage advice provided more accurate results, but lacked quality control and transparency.
The COVID-19 Symptom Tracker app uses user-reported symptoms to predict COVID-19 incidence hotspots up to a week in advance. Positive tests are often predicted by combinations of three or more symptoms, including fatigue and cough.
William Beksi, a UTA computer scientist, is developing topological methods to process large-scale 3D point cloud data, aiming to overcome limitations in robotic perception applications. This new approach uses persistent homology to extract unique insights into the structure of point cloud data.
A new study found that coercive measures like seclusion and mechanical restraint are commonly used in Finnish psychiatric care, with an average duration of nearly three days for seclusion. The use of coercive measures is more common than previously thought, highlighting the need for accurate data collection.
A mass study collaboration found that AI predictive models are not accurate in predicting life outcomes, particularly in the criminal justice system and social programs. The study used a large dataset of 13,000 data points from over 4,000 families, but even state-of-the-art modeling was not very accurate.
A collaborative study using machine learning techniques to predict six life outcomes for children, parents, and households found that even with high-quality data and state-of-the-art modeling, the predictive models were not very accurate. The study suggests that sociologists and data scientists should use caution in predictive modeling...
Researchers from North Carolina State University and the Army Research Office have developed a new model for how competing pieces of information spread in online social networks and IoT. The findings suggest that network size plays a significant role in displacing old data with new, accurate information.
Scientists developed a hybrid approach combining machine learning with nano-indentation techniques to improve the accuracy of material testing. The new method predicts samples' yield strength up to 20 times more accurately than existing methods.
Experts weigh the importance of balancing individual rights with innovation in regulating personal data. Singapore's Personal Data Protection Act (PDPA) aims to strike a balance between privacy protection and business interests.
A crowd-sourced study found that people who floss regularly have lower microbial diversity in their mouths than non-flossers, and adults with recent dental visits have less of the periodontal disease-causing oral pathogen. The study also uncovered a possible link between childhood obesity and periodontal disease.
Physicists at the Paul Scherrer Institute recorded a short 'film' of the three-dimensional magnetic structure inside a material with nanoscale resolution. This reveals intricate patterns and domain walls that could be used to pack data more tightly than current methods.
A new artificial neural network model has been developed to solve inverse problems, demonstrating accuracy comparable to the maximum entropy (MaxEnt) approach. The model's versatility and robustness against noisy data have been showcased in various tests, including recovering electron single-particle spectral densities.
Researchers at UC San Diego developed a new ultra-low power Wi-Fi radio that consumes just 28 microwatts of power, enabling IoT devices to communicate with existing Wi-Fi networks. The chip can transmit data at a rate of 2 megabits per second over a range of up to 21 meters.
A study by the University of Bath found that rugby's 'Big Brother' data culture can lead to heightened distrust, anxiety, and insecurity among players. Players feel judged on numbers rather than their true performance, leading to a loss of natural affinity and decision-making abilities.
Researchers propose a sampling regime inspired by geostatistics that includes regular and random intermediate points to capture fine-scale behaviors. This approach provides more accurate estimates of movement parameters than traditional regular sampling regimes.
Researchers propose a new computational analysis method that uses keystroke time series data to detect early stages of Parkinson's disease. The method, based on fuzzy recurrence plots, is less physically demanding than traditional testing methods and shows encouraging results.
The MIT-developed model uses anonymized credit card transactions and quarterly reports to predict company earnings, outperforming human analysts on 57% of predictions. The researchers' method tackles the 'small data' problem by combining noisy signals with traditional financial data.
A recent study from Duke University's Fuqua School of Business found that incumbent CEOs working with newly hired CFOs (co-opted CFOs) take home an average of 10% more compensation. This is attributed to the CEO's power over the CFO, who may be pressured to manage earnings to meet or exceed financial analyst targets, ultimately driving...
Researchers at the University of Helsinki developed an artificial intelligence-based algorithm to improve intensive care for patients with severe traumatic brain injuries. The algorithm accurately predicts patient mortality within 30 days, showing promising results in a proof-of-concept study.
A global health viewpoint emphasizes the need for accurate data to measure progress toward the United Nations Sustainable Development Goals. The lack of data in low- and middle-income countries, combined with inconsistent international standards, is hindering efforts to track progress.
A large-scale study of over 11,000 West Point cadets reveals that both cognitive and non-cognitive factors predict success outcomes. Grit is crucial during Beast Barracks, while cognitive ability is the strongest predictor of academic grades in the classroom. Physical ability also plays a key role in determining graduation rates.
Scientists at the University of Bonn develop an algorithm that completes and interprets LiDAR scans to enable self-driving cars to anticipate potential hazards. The system uses a dataset of superimposed point clouds to improve scene understanding, which can lead to significant improvements in autonomous driving safety.
A new protocol, On-Off Noise Power Communication (ONPC), extends Wi-Fi signal range by over 60 meters using software updates. The innovation requires no new hardware and can supplement existing Wi-Fi networks.
Scientists can study multiple regions in space simultaneously using sounding rockets equipped with sub-payloads and a distributed payload communications system. This technology allows for the transmission of data from multiple instruments at different altitudes, enabling new research opportunities.
Gun violence in America claims over 35,000 lives annually; UW researchers identify roadblocks to collecting and analyzing firearm data. Improved access to data could inform evidence-based policies to reduce gun violence.
Computational analysis identifies gaps in mosquito species distribution models, highlighting the need for improved data and models to advance disease understanding. Key regions with high uncertainty were found, particularly in borderland habitats and areas like Florida's panhandle.
A study analyzing prescribing data from over 8,000 English general practitioners reveals substantial variation in the adoption of new prescribing guidance, with some practices changing much later than others. This variation can lead to avoidable health service costs and poorer patient care.
The Rare Disease Cures Accelerator-Data and Analytics Platform (RDCA-DAP) aims to accelerate the development of treatments for rare diseases by providing data and analytics. The platform, launched by C-Path and NORD, will facilitate collaboration among key stakeholders, including patients, industry, and regulatory agencies.
Rice astroparticle physicist Christopher Tunnell leads a $1 million NSF-funded project to enhance data science techniques in physical sciences, aiming to push discovery past the tipping point. The study focuses on dark matter searches and employs probabilistic graphical models to improve measurements of particle interactions.
Researchers discovered 'survivor bias' in groundwater level monitoring, where wells with missing data are excluded, leading to incorrect conclusions. By re-including these wells, the study found that groundwater levels were actually declining in southern India.
Researchers trained an AI model using human-generated clickbait data, resulting in improved performance compared to other systems. The study found differences in headline creation between humans and machines, highlighting the need for high-quality training data to improve machine learning models.
Researchers developed an app called MedBlockChain that uses blockchain technology to prescribe optimal medicine doses for individual patients. The app also detects counterfeit products, providing a safer and more efficient way to manage pharmaceuticals.
A streamlined fee waiver process enabled about 73,000 low-income immigrants to become citizens each year, according to a new study from Stanford University's Immigration Policy Lab. The reform replaced a convoluted process with clear guidelines and a simple form, making it easier for those eligible to apply.
The SLAP microscope uses compressed measurements to scan large areas quickly, recording neurons' voltage spikes and neurotransmitter release. It has broken the speed limit of traditional two-photon microscopy, allowing scientists to capture millisecond-scale patterns in living brains.
MIT physicists develop technique to analyze open data from particle colliders, creating a geometric map of events based on degree of similarity, allowing them to quickly spot potential new physics. The researchers use an algorithm inspired by computer vision techniques to calculate the distance between point clouds in collision data.
Researchers use combinatorial synthesis and thin-film material libraries to accelerate discovery of new materials. Automated data analysis enables machine learning and artificial intelligence to aid the search for new materials.
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.
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.
Recent Chinese AI research has made significant breakthroughs in big data analysis, with a focus on developing algorithms to detect abnormal data regions and repair historical correlations. Researchers have also proposed using Doppler measurements to improve target tracking performance in noisy environments.
Recent US data shows a significant increase in suicide rates among adolescents (15-19) and young adults (20-24), with a 11.8 per 100,000 rate in 2017, up from 8 in 2000.