Recent work in signal propagation theory has shown that dropout limits the depth to which information can propagate through a neural network. In this paper, we investigate the effect of initialisation on training spee...
详细信息
The ability to learn and adapt in real time is a central feature of biological systems. Neuromorphic architectures demonstrating such versatility can greatly enhance our ability to efficiently process information at t...
详细信息
In this work, we consider multigoal-oriented error estimation for stationary fluid-structure interaction. The problem is formulated within a variational-monolithic setting using arbitrary Lagrangian-Eulerian coordinat...
详细信息
Global demand for data scientists and analysts is on a trajectory of sustained growth, underscoring the need to expand enrollment of students in data science and data analytics courses and programs. This article explo...
详细信息
Global demand for data scientists and analysts is on a trajectory of sustained growth, underscoring the need to expand enrollment of students in data science and data analytics courses and programs. This article explores diverse pathways to data science for various groups, including high school students, those in two-year colleges, and students in small colleges and minority-serving institutions. This article identifies opportunities and challenges associated with pursuing data science for students in these sectors, including factors such as building partnerships and awareness of the field and difficulties in establishing articulation agreements between institutions. The authors, representing various organizations participating in a panel discussion at the 2023 Data science Leadership Summit of the Academic Data science Alliance (ADSA), share insights into efforts to develop nontraditional pathways, serving as a resource for schools, colleges, and organizations contemplating creation of data scienceprograms. The panelists’ combined experience spans work with a broad range of organizations including ADSA, American mathematics Association of Two-Year Colleges, American Statistical Association, Data science for Everyone, the Department of Education, HBCU Data science Consortium, Mathematical Association of America, National Academies of science, Engineering, and Medicine, HBCU Data science Consortium, National science Foundation, and StatPREP.
Ensembles of thermostatically controlled loads (TCL) provide a significant demand response reserve for the system operator to balance power grids. However, this also results in the parasitic synchronization of individ...
详细信息
Features such as photon rings, jets, or hot spots can leave particular topological signatures in a black hole image. As such, topological data analysis can be used to characterize images resulting from high resolution...
详细信息
The K+ uptake system KtrAB is essential for bacterial survival in low K+ environments. The activity of KtrAB is regulated by nucleotides and Na+. Previous studies proposed a putative gating mechanism of KtrB regulated...
Recent advances in self-supervised learning enabled novel medical AI models, known as foundation models (FMs), offer great potential for better characterizing health from diverse biomedical data. Continuous glucose mo...
Recent advances in self-supervised learning enabled novel medical AI models, known as foundation models (FMs), offer great potential for better characterizing health from diverse biomedical data. Continuous glucose monitoring (CGM) provides rich, temporal data on glycemic patterns, but its full potential for predicting broader health outcomes remains underutilized. Here, we introduce GluFormer, a generative foundation model for CGM data designed to learn glycemic patterns and produce representations that reflect aspects of metabolic health. Trained on over 10 million CGM measurements from 10,812 adults, primarily without diabetes, GluFormer uses autoregressive token prediction to capture longitudinal glucose dynamics. We show that GluFormer generalizes to 19 external cohorts (n=6,044) spanning different ethnicities and ages, 5 different countries, 8 different CGM devices, and diverse pathophysiological states (e.g., prediabetes, type 1 & type 2 diabetes, gestational diabetes, and obesity). Our results suggest that GluFormer’s representations can improve upon current CGM metrics, such as the Glucose Management Indicator (GMI), for forecasting clinical measures. In individuals with prediabetes, GluFormer stratifies those likely to experience clinically significant increases in HbA1C% over a two-year period - outperforming baseline HbA1C%. In a longitudinal study of 580 adults with CGM data and 12-year follow-up, GluFormer identifies individuals at elevated risk of developing diabetes more effectively than blood HbA1C%, capturing 66% of all new-onset diabetes diagnoses in the top quartile versus 7% in the bottom quartile. Similarly, 69% of cardiovascular-death events occurred in the top quartile and none in the bottom quartile, indicating GluFormer-based stratification could capture additional risk signals not reflected by traditional glycemic metrics. We also show that CGM representations from pre-intervention periods in Randomized Clinical Trials (RCTs) outperform we
In single-cell RNA-sequencing (scRNA-seq), gene expression is assessed individually for each cell, allowing the investigation of developmental processes, such as embryogenesis and cellular differentiation and regenera...
详细信息
暂无评论