Background For the same body mass index (BMI) level, waist circumference (WC) is higher in more recent years. How this impacts diabetes and prediabetes prevalence in the United States and for different race/ethnic groups is unknown. We examined prevalence differences in diabetes and prediabetes by BMI over time, investigated whether estimates were attenuated after adjusting for waist circumference, and evaluated implications of these patterns on race/ethnic disparities in glycemic outcomes. Methods Data came from 12 614 participants aged 20 to 74 years from the National Health and Nutrition Examination Surveys (1988-1994 and 2007-2012). We estimated prevalence differences in diabetes and prediabetes by BMI over time in multivariable models. Relevant interactions evaluated race/ethnic differences. Results Among normal, overweight, and class I obese individuals, there were no significant differences in diabetes prevalence over time. However, among individuals with class II/III obesity, diabetes prevalence rose 7.6 percentage points in 2007-2012 vs 1988-1994. This estimate was partly attenuated after adjustment for mean waist circumference but not mean BMI. For prediabetes, prevalence was 10 to 13 percentage points higher over time at lower BMI values, with minimal attenuation after adjustment for WC. All patterns held within race/ethnic groups. Diabetes disparities among blacks and Mexican Americans relative to whites remained in both periods, regardless of BMI, and persisted after adjustment for WC. Conclusions Diabetes prevalence rose over time among individuals with class II/III obesity and may be partly due to increasing waist circumference. Anthropometric measures did not appear to account for temporal increases in prediabetes, nor did they attenuate race/ethnic disparities in diabetes. Reasons underlying these trends require further investigation.

Secular and race/ethnic trends in glycemic outcomes by BMI in US adults: The role of waist circumference. xxx

We examine whether having previously been self-employed is a negative signal on the job market. In a UK field experiment where two applications of otherwise equally qualified individuals were sent out in response to the same vacancies in human resource management, we find that entrepreneurs systematically receive fewer responses than non-entrepreneurs. Empirical studies that treat market wages as the opportunity cost of remaining self-employed are therefore likely to overestimate alternative earnings to entrepreneurship.

Self-Employed but Looking: A Labor Market Experiment

"Function calls as citations" refers to the ability to compose a novel research object by re-using an existing component, such as code or data, from a prior research object stored in an immutable and verifiable way. This idea originates from Balaji Srinivasan, who recently issued it as a challenge to builders in the DeSci space. In this preliminary report, we answer this challenge with a proof of concept. We created two Research Objects stored on an Open State Repository, each containing code components. A new Research Object calls functions from the code components of another Research Object. Both Research Objects are stored on an Open State Repository, and interoperate using decentralized persistent identifiers.

Function Calls as Citations

Transformer based models have provided significant performance improvements in monaural speech separation. However, there is still a performance gap compared to a recent proposed upper bound. The major limitation of the current dual-path Transformer models is the inefficient modelling of long-range elemental interactions and local feature patterns. In this work, we achieve the upper bound by proposing a gated single-head transformer architecture with convolution-augmented joint self-attentions, named MossFormer (Monaural speech separation TransFormer). To effectively solve the indirect elemental interactions across chunks in the dual-path architecture, MossFormer employs a joint local and global self-attention architecture that simultaneously performs a full-computation selfattention on local chunks and a linearised low-cost self-attention over the full sequence. The joint attention enables MossFormer model full-sequence elemental interaction directly. In addition, we employ a powerful attentive gating mechanism with simplified single-head self-attentions. Besides the attentive long-range modelling, we also augment MossFormer with convolutions for the position-wise local pattern modelling. As a consequence, MossFormer significantly outperforms the previous models and achieves the state-of-the-art results on WSJ0-2/3mix and WHAM!/WHAMR! benchmarks. Our model achieves the SI-SDRi upper bound of 21.2 dB on WSJ0-3mix and only 0.3 dB below the upper bound of 23.1 dB on WSJ0-2mix.

MOSSFORMER: PUSHING THE PERFORMANCE LIMIT OF MONAURAL SPEECH SEPARATION USING GATED SINGLE-HEAD TRANSFORMER WITH CONVOLUTION-AUGMENTED JOINT SELF-ATTENTIONS

Abstract We investigate the degree to which impact in science and technology is associated with surprising breakthroughs, the how those breakthroughs arise. Identifying the across the breakthroughs to requires models that distinguish impact from expected advances at scale. Drawing on tens of millions in research papers of patents and in life sciences, physical sciences science patented inventions, science using a hypergraph model science predicts realized combinations and and contents (article keywords) and contexts (cited journals), here we show and surprise technology terms surprising unexpected and and and and and and outsized technology (within across top 10% surprising citations). These that research emerge across, rather than within researchers or teams—most commonly when scientists that one field publish problem-solving results contexts an audience combinations contents distant field. Our approach characterizes of frontier from from from advances as of complex of drawn predicts high-dimensional embeddings of of of of contexts, research offers a measure a path-breaking a hypergraph contents surprise technology.

Attention Is All You Need

The BOLT-LMM software package computes statistics for association between phenotype and genotypes using a linear mixed model (LMM) [1]. By default, BOLT-LMM assumes a Bayesian mixture-of-normals prior for the random effect attributed to SNPs other than the one being tested. This model generalizes the standard "infinitesimal" mixed model used by existing mixed model association methods (e.g., EMMAX [2], FaST-LMM [3-6], GEMMA [7], GRAMMAR-Gamma [8], GCTA-LOCO [9]), providing an opportunity for increased power to detect associations while controlling false positives. Additionally, BOLT-LMM applies algorithmic advances to compute association statistics much faster than existing methods, both when using the Bayesian mixture model and when specialized to standard mixed model association.

BOLT-LMM v1.0 User Manual

tesig nde 10001

testing more, sdfgdsf gfdgfhgf atrestf 2325 65




testing4567912345fdadfs fdasfdssa fdsdfn jklnakm. sfdfdfdf hello

Deformable (2D or 3D) medical image registration is a challenging problem. Existing approaches assume that the underlying deformation is smooth. This smoothness assumption allows for solving the deformable registration at a coarse resolution and interpolate for finer resolutions. However, sliding of organs and breathing motion, exhibit discontinuities. We propose a discrete optimization approach to preserve these discontinuities. Solving continuous deformations using discrete optimization requires a fine distribution of the discrete labels. Coupled with the typical size of medical image datasets, this poses challenges to compute solutions efficiently. In this paper we present a practical, multi-scale formulation. We describe how discontinuities can be preserved, and how the optimization problem is solved. Results on synthetic 2D, and real 3D data show that we can well approximate the smoothness of continuous optimization, while accurately maintaining discontinuities.