Articles tagged with Data Sets
A new study from Aalto University found that global population datasets may underestimate rural populations by as much as 53% to 84%, revealing a systematic and profound extent of the bias. The findings have far-reaching consequences, suggesting that the needs of people living in rural areas have been under-represented in decision-maki...
A new study shows that climate change affects short-term precipitation events lasting hours by significantly increasing their frequency, while longer-term rainfall events lasting days are influenced by global weather phenomena like El Niéo. This distinction is crucial for assessing the risk of flooding in different regions.
Researchers have developed a novel technique to overcome the spurious correlations problem in AI by eliminating a small portion of the training data that contains hard-to-understand features. This approach improves performance even when conventional techniques are ineffective.
A recent USC study found that Black immigrants in the United States have a 9% higher likelihood of being uninsured compared to U.S.-born white adults, despite having a 70.2% employment rate. The disparity is attributed to wage penalties and systemic barriers related to race and immigration status.
FAIR² Data Management leverages AI-assisted curation to structure research data for publication, enhancing research efficiency and reproducibility. The platform transforms data into a tool for discovery, tackling pressing challenges in global health, sustainability, and scientific innovation.
Researchers at Trinity College Dublin have discovered a biomarker panel that can accurately identify patients at risk of developing pancreatic cancer, offering new hope for early detection and improved treatment. The study uses a combination of blood and fluid analysis to distinguish between low- and high-risk patients.
A new standard dataset defines 16 clinical and patient-reported outcome measures for adults with heart valve disease. The International Consortium for Health Outcomes Measurement (ICHOM) developed the globally inclusive standard dataset.
The release of a unique Type Ia Supernovae dataset has significant implications for cosmologists measuring the universe's expansion history. The dataset, comprising 3628 supernovae, provides unprecedented precision and accuracy in exploring the properties of these events.
Researchers generated a dataset of water flow beneath the entire Antarctic Ice Sheet, improving projections of sea level rise. The model predicts higher degrees of glacier melt and mass loss by the end of the century.
A new method developed by Osaka Metropolitan University accurately predicts housing prices in Osaka City, with neighborhood perception being a key factor. The approach achieves nearly 75% accuracy by combining existing property data with machine-learning-processed street view images.
A new study found that hotter and colder regions on a star's surface can distort our interpretations of planets, particularly when looking at dips in starlight. This distortion can lead to misinterpretation of features such as planet size, temperature, and atmospheric composition.
InsectNet uses machine learning to identify over 2,500 insect species at 96% accuracy, providing critical information for farmers and researchers. The app can be fine-tuned for specific regions, making it useful for agricultural challenges worldwide.
The International Society for Health and Lung Transplantation (ISHLT) has re-launched the Global IMACS Registry to collect data on patients with durable mechanical circulatory support devices. The registry aims to advance MCS therapy and improve patient outcomes globally.
Two comprehensive datasets from the Gabriella Miller Kids First Pediatric Research Program explore childhood cancers and congenital disorders. The new datasets aim to identify genetic causes and links between these diseases in children, ultimately supporting the development of improved treatments.
A new computational tool, PsiPartition, simplifies genetic data analysis for evolutionary biology, allowing researchers to efficiently study species relationships. The novel method improves both computational efficiency and accuracy of phylogenetic trees.
Researchers at University of Birmingham have discovered three new protein biomarkers TFF3, LCN2, and CEACAM5 that show strong predictive potential for colorectal cancer. These biomarkers are linked to cell adhesion and inflammation, processes closely associated with cancer development.
A new 6D pose dataset has been introduced, providing high-quality RGB and depth images with annotated 6D pose data. This dataset achieves state-of-the-art accuracy rates of 97.05% and 98.09% for robotic grasping and automation applications.
A team led by the University of Tokyo has created a nearly 20-year-long dataset of the entire atmosphere, enabling new research on previously difficult-to-study regions. The dataset spans multiple levels of the atmosphere from ground level to the lower edge of space and could improve climate modeling and seasonal weather forecasting.
The UK Biobank has launched a groundbreaking study to measure over 5,400 proteins in 600,000 samples, offering unprecedented insights into disease development and progression. This massive dataset will enable researchers to explore how changes in protein levels influence diseases and develop personalized treatments.
Researchers at the University of New Hampshire developed an AI-powered algorithm to categorize over 706 million aurora images from NASA's THEMIS data set. This labeled database can help scientists better understand and forecast geomagnetic storms that disrupt vital communications and security infrastructure.
A large UK-wide study has identified specific clusters of long-term health conditions associated with lower quality of life, particularly those linked to chronic pain, depression, and cardiovascular disease. The findings highlight the need for tailored treatment approaches for patients with multiple long-term health conditions.
Researchers reveal that contradictory findings on territorial self-governance and ethnic civil war risk are primarily due to variations in measurement definitions. The study highlights the importance of considering the nuances of data choice in understanding conflict resolution.
A new method called Annotatability helps identify mismatches in cell annotations and better characterizes biological data structures. This approach enables more precise downstream analysis of biological signals, capturing cellular communities associated with target signals.
A new book analyzes data from 48 US states from 2008 to 2016, finding that election costs vary significantly across states and regions. The authors also found that trust in election spending is associated with how much is spent, and that a one-size-fits-all approach to legislation does not apply to election administration.
Researchers from Florida Atlantic University developed an AI system that accurately detects and tracks American Sign Language alphabet gestures using computer vision. The model achieved a detection process with high accuracy, recognizing complex hand gestures with an F1 score of 99% and mAP of 98%.
Researchers at Tokyo University of Science have developed a new method called black-box forgetting, which enables selective removal of unnecessary information from large pre-trained AI models. This approach enhances model efficiency and improves privacy by reducing computational resources and information leakage.
Insilico Medicine has successfully integrated cutting-edge generative AI models into its platform, achieving significant increases in accuracy and speed. The company's AI-powered solutions have shown promising results in preclinical candidates and Phase II clinical trials.
The partnership integrates Answer ALS’s Neuromine Data Portal with the AD Data Initiative's AD Workbench, providing a unified platform for data sharing and analysis. This collaboration enables integrative analysis of ALS and Alzheimer’s data, streamlining efforts and enhancing research impact.
Researchers from NUS and A*STAR have discovered a connection between the regulation of alternative splicing in different cell types and the predisposition to autoimmune diseases. The study used a population-scale single-cell gene expression profiling dataset to analyze splicing events specific to particular cell types, revealing ancest...
The latest report in the State of Open Data series provides quantitative analysis on the growth of open data sharing globally. Key findings show a strong growth in papers linking to data from universities, and policies for open data sharing are now more consistent globally.
The study discovered significant alterations in the region's state of stress and deformation following the 1975 Kalapana earthquake. The researchers found that Kīlauea's south flank experienced greater displacement prior to the earthquake, pointing to changes in mechanical properties influencing seismic activity.
The Polymathic AI team has released two massive datasets for training artificial intelligence models to find and exploit transferable knowledge between seemingly disparate fields. The datasets include data from dozens of sources, covering astrophysics, biology, acoustics, chemistry, fluid dynamics, and more.
A novel framework for retrieving concise entailing legal articles has been proposed, achieving state-of-the-art results across two datasets. The Retrieve–Revise–Refine framework combines small and large language models to improve precision while limiting recall loss.
The University of Rhode Island-based report card gives failing grades to over half the world's countries, with a median score of 52. Democratic countries tend to have better human rights records, but exceptions exist. The US scored a 62.5, ranking 66th, despite being a wealthy democratic country.
The National Center for Supercomputing Applications (NCSA) has received the Readers' Choice Award: Best HPC Collaboration and Editors' Choice: Best Use of HPC in Physical Sciences. This is the 14th consecutive year NCSA has been honored with an HPCwire award.
Researchers found that membership inference attacks on large language models (LLMs) are not effective in measuring information exposure risks. The common method used to test LLM leaks suffers from ambiguity due to the fluidity of language, making it difficult to define a representative set of non-member candidates.
A panel of bioethicists and experts emphasize that human accountability is crucial for healthcare decisions made by AI. The importance of diverse data sets was also stressed to avoid biases in AI-enabled medical technologies. Experts stress the need for collective liability among developers, programmers, and data scientists.
A recent machine learning study suggests that the association between gut bacteria and disease may be overstated. Instead, changes in microbial load were found to be a key factor in the presence of disease-associated microbial species. This discovery challenges current understanding of the gut microbiome's role in disease etiology.
Researchers at Lehigh University are using advanced algorithms and cross-domain data to help cities predict human movement patterns, enabling better planning and preparedness for events and emergencies. The model will account for variations in data streams from different sources, such as cell towers, GPS, and financial transactions.
The University of Washington School of Medicine has released a flagship AI-ready dataset for a type 2 diabetes study, featuring a diverse range of participants and novel measures. The dataset includes environmental sensor data, survey responses, eye-imaging scans, and traditional biologic variables.
Researchers developed a multimodal dataset, TimelyTale, to gather passenger-specific sensor data for context-relevant explanations. The approach effectively identified the timing and frequency of passenger demands for explanations, enabling the creation of a machine-learning model to predict the best time for providing an explanation.
Researchers at the University of Liverpool developed AI-driven mobile robots that can perform exploratory chemistry research tasks faster and more efficiently than humans. The robots use AI logic to make decisions, processing analytical datasets in real-time to determine the next steps in chemical synthesis.
Researchers at Radboud University Medical Center found that an extra year of education does not protect against brain aging and has no effect on brain structure. Despite positive correlations between education and cognitive benefits, the study suggests caution in assigning causation when only correlation is observed.
A new study by the Complexity Science Hub finds that firms with higher levels of information consumption outperform their peers financially and are more innovative. The research reveals an 'economy of scale' in news consumption, where larger firms read a greater number of unique pieces of news.
A new genetic analysis method called Genomic Informational Theory (GIFT) has been developed to extract more precise data than previously used methods. GIFT is capable of analyzing large datasets and extracting novel information that was previously unavailable through genome-wide association studies (GWASs).
A novel collaborative framework integrates semi-supervised learning techniques to improve MRI segmentation accuracy, even with limited labeled data. The approach achieves high Dice scores and demonstrates its potential for practical clinical application.
MIT researchers developed a versatile technique that combines diverse data from various sources into a shared language for generative AI models. This approach outperformed traditional techniques by 20% in simulation and real-world experiments.
Researchers are combining large datasets to understand the impact of environmental factors on mental health and develop new ways to reduce the burden. A 4-step approach is being used to integrate complex data across many studies, resulting in a detailed perspective of individuals within their environment.
The COP16 policy brief recommends implementing FAIR data principles to enhance biodiversity monitoring. This approach enables standardization of indicators, streamlined reporting, and improved global tracking.
A new Multi-task Learning (MTL) model detects 85% of abusive posts originating from right-leaning individuals on social media platforms.
The two-year study aims to explore biases in AI systems and develop a 'human-in-the-loop' framework for quality data discovery. It will investigate how humans can be involved as labelers, prompters, and validators to improve data sets and user interfaces.
A study reveals that shrinking Japanese cities are often medium-sized or small, and that urban policies should be tailored to their specific needs. The research found correlations between population changes and social, economic, and urban-planning factors in cities of varying sizes.
A research team led by Maria Glymour will investigate four modifiable risk factors for ADRD: lifetime alcohol use, depression, vision and hearing impairments, and social isolation. The project aims to provide comprehensive evidence on how these risk factors contribute to ADRD risk.
Scientists analyzed millions of tweets to detect early warning signs of PTSD in COVID-19 survivors, achieving an accuracy rate of 83.29%. The study highlights the potential for machine learning techniques to identify individuals at risk of PTSD through social media data.
A research team developed a computational workflow for analyzing large data sets in metabolomics, speeding up the process to capture chemical profiles of coastal environments. The tool, accessible to researchers worldwide, highlights potential sources of pollution and enables statistical insights within minutes.
A Concordia-led team developed a framework that enables crowdsourced deep reinforcement learning as a service, using blockchain technology. This allows smaller organizations to access complex AI tasks previously out of reach, reducing costs and risk.
Researchers identified known genetic variants for Alzheimer’s disease as risk factors for all-cause dementia and vascular dementia. The study also found a substantial genetic overlap with vascular dementia and cerebral small-vessel disease.
A novel approach to overcome limitations of traditional methods, NeuPh uses local conditional neural fields to reconstruct high-resolution phase information from low-resolution measurements. It provides robust resolution enhancement and outperforms existing models in accuracy.
Researchers found that control of most genes doesn't deteriorate with age, but coordination between cellular processes becomes less effective. The study suggests a more complex approach to understanding aging is needed, analyzing all genes simultaneously and their protein interactions.
Researchers developed a tool to improve data transparency in large language models, enabling practitioners to find suitable datasets for their models. The tool, Data Provenance Explorer, automatically generates summaries of dataset creators, sources, licenses, and allowable uses.