Showing papers by "University of Colorado Colorado Springs published in 2018"
TL;DR: This study assesses the state-of-the-art machine learning methods used for brain tumor image analysis in mpMRI scans, during the last seven instances of the International Brain Tumor Segmentation (BraTS) challenge, i.e., 2012-2018, and investigates the challenge of identifying the best ML algorithms for each of these tasks.
Abstract: Gliomas are the most common primary brain malignancies, with different degrees of aggressiveness, variable prognosis and various heterogeneous histologic sub-regions, i.e., peritumoral edematous/invaded tissue, necrotic core, active and non-enhancing core. This intrinsic heterogeneity is also portrayed in their radio-phenotype, as their sub-regions are depicted by varying intensity profiles disseminated across multi-parametric magnetic resonance imaging (mpMRI) scans, reflecting varying biological properties. Their heterogeneous shape, extent, and location are some of the factors that make these tumors difficult to resect, and in some cases inoperable. The amount of resected tumoris a factor also considered in longitudinal scans, when evaluating the apparent tumor for potential diagnosis of progression. Furthermore, there is mounting evidence that accurate segmentation of the various tumor sub-regions can offer the basis for quantitative image analysis towards prediction of patient overall survival. This study assesses thestate-of-the-art machine learning (ML) methods used for brain tumor image analysis in mpMRI scans, during the last seven instances of the International Brain Tumor Segmentation (BraTS) challenge, i.e., 2012-2018. Specifically, we focus on i) evaluating segmentations of the various glioma sub-regions in pre-operative mpMRI scans, ii) assessing potential tumor progression by virtue of longitudinal growth of tumor sub-regions, beyond use of the RECIST/RANO criteria, and iii) predicting the overall survival from pre-operative mpMRI scans of patients that underwent gross tota lresection. Finally, we investigate the challenge of identifying the best ML algorithms for each of these tasks, considering that apart from being diverse on each instance of the challenge, the multi-institutional mpMRI BraTS dataset has also been a continuously evolving/growing dataset.
1,165 citations
TL;DR: The Extreme Value Machine (EVM) is a novel, theoretically sound classifier that has a well-grounded interpretation derived from statistical Extreme Value Theory (EVT), and is the first classifier to be able to perform nonlinear kernel-free variable bandwidth incremental learning.
Abstract: It is often desirable to be able to recognize when inputs to a recognition function learned in a supervised manner correspond to classes unseen at training time. With this ability, new class labels could be assigned to these inputs by a human operator, allowing them to be incorporated into the recognition function—ideally under an efficient incremental update mechanism. While good algorithms that assume inputs from a fixed set of classes exist, e.g. , artificial neural networks and kernel machines, it is not immediately obvious how to extend them to perform incremental learning in the presence of unknown query classes. Existing algorithms take little to no distributional information into account when learning recognition functions and lack a strong theoretical foundation. We address this gap by formulating a novel, theoretically sound classifier—the Extreme Value Machine (EVM). The EVM has a well-grounded interpretation derived from statistical Extreme Value Theory (EVT), and is the first classifier to be able to perform nonlinear kernel-free variable bandwidth incremental learning. Compared to other classifiers in the same deep network derived feature space, the EVM is accurate and efficient on an established benchmark partition of the ImageNet dataset.
272 citations
Proceedings Article•
01 Jan 2018TL;DR: This paper introduces a new evaluation metric that focuses on comparing the performance of multiple approaches in scenarios where such unseen classes or unknowns are encountered, and develops novel loss functions that train networks using negative samples from some classes.
Abstract: Agnostophobia, the fear of the unknown, can be experienced by deep learning engineers while applying their networks to real-world applications. Unfortunately, network behavior is not well defined for inputs far from a networks training set. In an uncontrolled environment, networks face many instances that are not of interest to them and have to be rejected in order to avoid a false positive. This problem has previously been tackled by researchers by either a) thresholding softmax, which by construction cannot return "none of the known classes", or b) using an additional background or garbage class. In this paper, we show that both of these approaches help, but are generally insufficient when previously unseen classes are encountered. We also introduce a new evaluation metric that focuses on comparing the performance of multiple approaches in scenarios where such unseen classes or unknowns are encountered. Our major contributions are simple yet effective Entropic Open-Set and Objectosphere losses that train networks using negative samples from some classes. These novel losses are designed to maximize entropy for unknown inputs while increasing separation in deep feature space by modifying magnitudes of known and unknown samples. Experiments on networks trained to classify classes from MNIST and CIFAR-10 show that our novel loss functions are significantly better at dealing with unknown inputs from datasets such as Devanagari, NotMNIST, CIFAR-100 and SVHN.
187 citations
TL;DR: In this article, the authors examined the relationship between both micro (intrinsic and extrinsic motivation) and molar (team climate) variables with manager-rated creativity of R&D employees.
129 citations
Posted Content•
TL;DR: An introduction to the field and a quick overview of deep learning architectures and methods is provided and a discussion of the current state of the art is provided along with recommendations for future research in the field.
Abstract: Over the last several years, the field of natural language processing has been propelled forward by an explosion in the use of deep learning models. This survey provides a brief introduction to the field and a quick overview of deep learning architectures and methods. It then sifts through the plethora of recent studies and summarizes a large assortment of relevant contributions. Analyzed research areas include several core linguistic processing issues in addition to a number of applications of computational linguistics. A discussion of the current state of the art is then provided along with recommendations for future research in the field.
129 citations
TL;DR: Results support emerging research showing sleep quality is a risk factor for negative maternal affect in the postpartum period, and assessment of maternal sleep hygiene is worth consideration as a component of identifying women at risk for post partum depression and anxiety.
Abstract: This study evaluated the relationship between sleep quality and symptoms of depression and anxiety in women studied in pregnancy and postpartum. Scores on standardized measures of sleep (PSQI) at 6 months postpartum, and symptoms of anxiety and depression (OASIS, the PHQ9, and EPDS) were assessed by structured interviews in 116 women in pregnancy and/or postpartum. Poor sleep quality was significantly associated with greater symptoms of depression and anxiety. Women who had significantly higher OASIS (anxiety) scores (β = .530, p < .001), PHQ9 (depression) scores (β = .496, p < .001), and EPDS (postpartum depression and anxiety) scores (β = .585, p < .001) also had elevated total PSQI scores after adjustment for covariates, including prenatal depression and anxiety scores. Though inferences about causality are not feasible, these results support emerging research showing sleep quality is a risk factor for negative maternal affect in the postpartum period. Assessment of maternal sleep hygiene is worth consideration as a component of identifying women at risk for postpartum depression and anxiety.
124 citations
TL;DR: In a large, multi-institutional series of patients with liver metastasis treated with SBRT, reasonable LC and OS was observed and depended on dose and tumor volume, while OS varied by primary tumor.
Abstract: Stereotactic body radiotherapy (SBRT) is an emerging treatment option for liver metastases in patients unsuitable for surgery. We investigated factors associated with clinical outcomes for liver metastases treated with SBRT from a multi-center, international patient registry. Patients with liver metastases treated with SBRT were identified in the RSSearch® Patient Registry. Patient, tumor and treatment characteristics associated with treatment outcomes were assessed. Dose fractionations were normalized to BED10. Overall survival (OS) and local control (LC) were evaluated using Kaplan Meier analysis and log-rank test. The study included 427 patients with 568 liver metastases from 25 academic and community-based centers. Median age was 67 years (31–91 years). Colorectal adenocarcinoma (CRC) was the most common primary cancer. 73% of patients received prior chemotherapy. Median tumor volume was 40 cm3 (1.6–877 cm3), median SBRT dose was 45 Gy (12–60 Gy) delivered in a median of 3 fractions [1–5]. At a median follow-up of 14 months (1–91 months) the median overall survival (OS) was 22 months. Median OS was greater for patients with CRC (27 mo), breast (21 mo) and gynecological (25 mo) metastases compared to lung (10 mo), other gastro-intestinal (GI) (18 mo) and pancreatic (6 mo) primaries (p < 0.0001). Smaller tumor volumes (< 40 cm3) correlated with improved OS (25 months vs 15 months p = 0.0014). BED10 ≥ 100 Gy was also associated with improved OS (27 months vs 15 months p < 0.0001). Local control (LC) was evaluable in 430 liver metastases from 324 patients. Two-year LC rates was better for BED10 ≥ 100 Gy (77.2% vs 59.6%) and the median LC was better for tumors < 40 cm3 (52 vs 39 months). There was no difference in LC based on histology of the primary tumor. In a large, multi-institutional series of patients with liver metastasis treated with SBRT, reasonable LC and OS was observed. OS and LC depended on dose and tumor volume, while OS varied by primary tumor. Future prospective trials on the role of SBRT for liver metastasis from different primaries in the setting of multidisciplinary management including systemic therapy, is warranted. Clinicaltrials.gov: NCT01885299
.
122 citations
TL;DR: It is shown that diverse, low-level immune activity predicts reduced childhood growth over periods of competing energy use ranging from 1 wk to 20 mo, and that modest body fat stores protect children from the particularly detrimental impact of acute inflammation on growth.
Abstract: Immune function is an energetically costly physiological activity that potentially diverts calories away from less immediately essential life tasks. Among developing organisms, the allocation of energy toward immune function may lead to tradeoffs with physical growth, particularly in high-pathogen, low-resource environments. The present study tests this hypothesis across diverse timeframes, branches of immunity, and conditions of energy availability among humans. Using a prospective mixed-longitudinal design, we collected anthropometric and blood immune biomarker data from 261 Amazonian forager-horticulturalist Shuar children (age 4-11 y old). This strategy provided baseline measures of participant stature, s.c. body fat, and humoral and cell-mediated immune activity as well as subsample longitudinal measures of linear growth (1 wk, 3 mo, 20 mo) and acute inflammation. Multilevel analyses demonstrate consistent negative effects of immune function on growth, with children experiencing up to 49% growth reduction during periods of mildly elevated immune activity. The direct energetic nature of these relationships is indicated by (i) the manifestation of biomarker-specific negative immune effects only when examining growth over timeframes capturing active competition for energetic resources, (ii) the exaggerated impact of particularly costly inflammation on growth, and (iii) the ability of children with greater levels of body fat (i.e., energy reserves) to completely avoid the growth-inhibiting effects of acute inflammation. These findings provide evidence for immunologically and temporally diverse body fat-dependent tradeoffs between immune function and growth during childhood. We discuss the implications of this work for understanding human developmental energetics and the biological mechanisms regulating variation in human ontogeny, life history, and health.
117 citations
TL;DR: Workplace spirituality has been conceptualized as offering new insights into how individuals experience a deeper level of intrinsic work motivation and engagement as mentioned in this paper, and workplace spirituality has a direct effect on employee engagement and intention to stay in a study of 292 employees in a U.S. hospitality organization.
113 citations
TL;DR: In this article, the authors focus on the evolutionary-functional level of analysis, outlining hypotheses capable of explaining why women have higher levels of disgust than men and present four hypotheses for sexual disgust and six for pathogen disgust, along with testable predictions.
Abstract: Women have consistently higher levels of disgust than men. This sex difference is substantial in magnitude, highly replicable, emerges with diverse assessment methods, and affects a wide array of outcomes—including job selection, mate choice, food aversions, and psychological disorders. Despite the importance of this far-reaching sex difference, sound theoretical explanations have lagged behind the empirical discoveries. In this article, we focus on the evolutionary-functional level of analysis, outlining hypotheses capable of explaining why women have higher levels of disgust than men. We present four hypotheses for sexual disgust and six for pathogen disgust, along with testable predictions. Discussion focuses on additional new hypotheses and on future research capable of adjudicating among these competing, but not mutually exclusive, hypotheses.
110 citations
TL;DR: To perform better in the WSTB and job-specific tasks, developing upper-body strength and aerobic fitness may be beneficial.
Abstract: This study determined relationships between an agency-specific fitness test battery (PT500), and a work sample test battery (WSTB) in law enforcement recruits. Retrospective analysis on 219 males and 34 females from one agency was conducted. The PT500 comprised: push-ups, sit-ups, and mountain climbers in 120 s; pull-ups; and 201 m and 2.4 km runs. The WSTB comprised: 99 yard (90.53 m) obstacle course (99OC); body drag (BD) with a 165 pound (75 kg) dummy; 6 foot (1.83 m) chain link fence (CLF) and solid wall (SW) climb; and 500 yard (457.2 m) run (500R). Partial correlations, controlling for sex, calculated PT500 and WSTB relationships (p < 0.05). Stepwise regression determined whether fitness predicted WSTB performance. The 500R related to all PT500 assessments (r range = −0.127–0.574), 99OC related to all bar push-ups and mountain climbers, and BD related to none. The CLF related to sit-ups, pull-ups, and 2.4 km run; SW related to mountain climbers, pull-ups, and 2.4 km run (r range = −0.127–−0.315). Push-ups, pull-ups, and 2.4 km run were involved in predictive relationships for 99OC, CLF, SW, and 500R (r2 range = 0.217–0.500). To perform better in the WSTB and job-specific tasks, developing upper-body strength and aerobic fitness may be beneficial.
TL;DR: The results suggest that education is an indicator of cognitive reserve in individuals with low levels of brain degeneration, but the protective effect of higher education is rapidly depleted asbrain degeneration progresses.
18 Jun 2018
TL;DR: In this article, the authors use dilated filters to construct an aggregation module in a multicolumn convolutional neural network for perspective-free counting, which achieves state-of-the-art performance on many benchmark datasets.
Abstract: We propose the use of dilated filters to construct an aggregation module in a multicolumn convolutional neural network for perspective-free counting. Counting is a common problem in computer vision (e.g. traffic on the street or pedestrians in a crowd). Modern approaches to the counting problem involve the production of a density map via regression whose integral is equal to the number of objects in the image. However, objects in the image can occur at different scales (e.g. due to perspective effects) which can make it difficult for a learning agent to learn the proper density map. While the use of multiple columns to extract multiscale information from images has been shown before, our approach aggregates the multiscale information gathered by the multicolumn convolutional neural network to improve performance. Our experiments show that our proposed network outperforms the state-of-the-art on many benchmark datasets, and also that using our aggregation module in combination with a higher number of columns is beneficial for multiscale counting.
TL;DR: The reactions of Ga(i-Bu)3 with the dehydrated and partially dehydroxylated surfaces of alumina (Al2O3-500) and silica (SiO2-700) were studied by IR, high field solid-state NMR and EXAFS spectroscopies, as well as elemental analysis as mentioned in this paper.
Abstract: The reactions of Ga(i-Bu)3 (i-Bu = CH2CH(CH3)2) with the dehydrated and partially dehydroxylated surfaces of alumina (Al2O3–500) and silica (SiO2–700) were studied by IR, high field solid-state NMR and EXAFS spectroscopies, as well as elemental analysis. Grafting onto Al2O3–500 occurs selectively by protonolysis at individual surface hydroxyl groups, resulting in the formation of mononuclear [(≡AlO)Ga(i-Bu)2L] (L = surface oxygen) sites as the major surface organometallic entities. Conversely, grafting on silica affords dinuclear species [(≡SiO)2Ga2(i-Bu)3] by a combination of protonolysis and isobutyl group transfer to Si. Further evidence for the difference in nuclearity was obtained by analysis of the WT-EXAFS. The mononuclear alumina-supported Ga sites show much higher activity in propane dehydrogenation than their dinuclear silica-supported counterparts. The propane dehydrogenation reaction may require the presence of Al–O–Ga bonds to promote heterolytic C–H bond activation. Comparisons with benchmar...
TL;DR: In this article, the authors demonstrate that the application of HR policies for mid-level marketing managers vary significantly both between firms pursuing alternative business strategies (i.e., Prospectors, Analyzers, Low Cost Defenders and Differentiated Defenders) and within each of those business strategy types by the type of marketing strategy adopted.
TL;DR: It is found that the mode blocking temperature is at a lower temperature than the peak in the zero-field-cooled magnetisation versus temperature curve, in agreement with experiment and previous rate-equation simulations, but in contrast to the assumption many researchers use to analyse experimental data.
Abstract: We consider the probability of a magnetic nanoparticle to flip its magnetisation near the blocking temperature, and use this to develop quasi-analytic expressions for the zero-field-cooled and field-cooled magnetisation, which go beyond the usual critical energy barrier approach to the superparamagnetic transition. The particles in the assembly are assumed to have random alignment of easy axes, and to not interact. We consider all particles to be of the same size and then extend the theory to treat polydisperse systems of particles. In particular, we find that the mode blocking temperature is at a lower temperature than the peak in the zero-field-cooled magnetisation versus temperature curve, in agreement with experiment and previous rate-equation simulations, but in contrast to the assumption many researchers use to analyse experimental data. We show that the quasi-analytic expressions agree with Monte Carlo simulation results but have the advantage of being very quick to use to fit data. We also give an example of fitting experimental data and extracting the anisotropy energy density K.
TL;DR: In this paper, the authors identify challenges, disruptions, and contradictions as they occur across schools engaged in implementing technology-mediated personalized learning, and examine some of the structural and contextual sources of these disruptions and contradictions.
Abstract: In the current educational context, school models that leverage technology to personalize instruction have proliferated, as has student enrollment in, and funding of, such school models. However, even the best laid plans are subject to challenges in design and practice, particularly in the dynamic context of a school. In this collective case study, we identify challenges, disruptions, and contradictions as they occur across schools engaged in implementing technology-mediated personalized learning. Using cultural historical activity theory—a theoretical framework concerned with the individual and contextual factors influencing school change—to frame the analysis, we also examine some of the structural and contextual sources of these disruptions and contradictions. Our findings enable us to offer recommendations for policymakers and for practitioners engaged in implementing personalized learning models, as well as directions for future research on this topic.
TL;DR: Age-specific recommendations for basketball participation are provided that aim to promote a healthy and positive experience for youth basketball players.
Abstract: Participation in sports offers both short-term and long-term physical and psychosocial benefits for children and adolescents. However, an overemphasis on competitive success in youth sports may limit the benefits of participation, and could increase the risk of injury, burnout, and disengagement from physical activity. The National Basketball Association and USA Basketball recently assembled a group of leading experts to share their applied research and practices to address these issues. This review includes the group’s analysis of the existing body of research regarding youth sports participation and the related health, performance, and psychosocial outcomes. Based upon this, age-specific recommendations for basketball participation are provided that aim to promote a healthy and positive experience for youth basketball players.
11 Sep 2018
TL;DR: A theoretically based model of engagement is proposed that considers the unique challenges of trauma recovery and is intended to highlight the challenges of engagement research including its definition, measurement, and modeling.
Abstract: Exposure to traumatic events is extremely common with nearly 75% reported to have experienced one or more traumatic events worldwide. A significant number of those exposed will develop posttraumatic stress disorder (PTSD) along with depression, anxiety, and substance use disorders. Globally, trauma-related mental health disorders are the leading cause of global disability burden, and many of these disorders are caused, or worsened, by exposure to wars, natural and human-caused disasters, and other traumatic events. Significant barriers to treatment exist including logistical, geographical, financial, stigma, and other attitudinal challenges. One opportune approach to overcoming these barriers is the provision of mental health interventions via technology that can be readily standardized for wide dissemination of evidence-based care. However, engagement with technology-based interventions is a concern and limited participation and high attrition rates are common. This may be especially true for trauma survivors who often experience symptoms of avoidance and hyperarousal. Engagement is regarded as an essential component of intervention efficacy and has been demonstrated to be associated with more positive clinical outcomes, yet theoretically based research in this area is sparse. This review focuses on the complex issue of engagement with digital health interventions (DHIs). Specifically, we review the conceptualization and measurement of engagement, predictors of engagement, and importantly, the relationship of engagement with intervention effectiveness. Finally, a theoretically based model of engagement is proposed that considers the unique challenges of trauma recovery. This review is not intended to provide a systematic or exhaustive set of recommendations, rather it is intended to highlight the challenges of engagement research including its definition, measurement, and modeling. Future engagement research that includes valid and reliable measures of engagement will enable consistent exploration of engagement predictors that can then inform methods for increasing engagement and, ultimately, intervention effectiveness.
TL;DR: A new software, called tsscds2018, has been developed to discover reaction mechanisms and solve the kinetics in a fully automated fashion and employs algorithms based on Graph Theory to find transition state geometries from accelerated semiempirical dynamics simulations carried out with MOPAC2016.
Abstract: A new software, called tsscds2018, has been developed to discover reaction mechanisms and solve the kinetics in a fully automated fashion. The program employs algorithms based on Graph Theory to find transition state (TS) geometries from accelerated semiempirical dynamics simulations carried out with MOPAC2016. Then, the TSs are connected to the corresponding minima and the reaction network is obtained. Kinetic data like populations vs time or the abundancies of each product can also be obtained with our program thanks to a Kinetic Monte Carlo routine. Highly accurate ab initio potential energy diagrams and kinetics can also be obtained using an interface with Gaussian09. The source code is available on the following site: http://forge.cesga.es/wiki/g/tsscds/HomePage © 2018 Wiley Periodicals, Inc.
TL;DR: Performance of both power and agility was shown to decrease when tactical load was added to the participants, suggesting that the increase in weight carried by tactical officers may put this population at risk of injury or fatality in the line of duty.
Abstract: The current literature suggests that load carriage can impact on a tactical officer's mobility, and that survival in the field may rely on the officer's mobility. The ability for humans to generate power and agility is critical for performance of the high-intensity movements required in the field of duty. The aims of this review were to critically examine the literature investigating the impacts of load carriage on measures of power and agility and to synthesize the findings. The authors completed a search of the literature using key search terms in four databases. After relevant studies were located using strict inclusion and exclusion criteria, the studies were critically appraised using the Downs and Black Checklist and relevant data were extracted and tabled. Fourteen studies were deemed relevant for this review, ranging in percentage quality scores from 42.85% to 71.43%. Outcome measures used in these studies to indicate levels of power and agility included short-distance sprints, vertical jumps, and agility runs, among others. Performance of both power and agility was shown to decrease when tactical load was added to the participants. This suggests that the increase in weight carried by tactical officers may put this population at risk of injury or fatality in the line of duty.
TL;DR: A robust detailed national and international prospective database was recommended by the International Liaison Committee on Resuscitation in 2015 to facilitate further research unique to cardiac arrest during pregnancy that will produce optimal resuscitation techniques for maternal cardiac arrest.
TL;DR: Male recruits demonstrated superior performance across all power tests compared with the female recruits, and female recruits aged 35+ years of age may be lacking in upper- and lower-body power.
Abstract: Lockie, RG, Dawes, JJ, Orr, RM, Stierli, M, Dulla, JM, and Orjalo, AJ. Analysis of the effects of sex and age on upper- and lower-body power for law enforcement agency recruits before academy training. J Strength Cond Res 32(7): 1968–1974, 2018—Power is an important characteristic for law en
TL;DR: In this paper, poly(methyl methacrylate) (PMMA) was added to aluminum/poly(vinylidene fluoride) (Al/PVDF) energetic blends to enhance melt flow rate and adhesion in a fused deposition modeling (FDM) manufacturing scenario.
Abstract: Poly(methyl methacrylate) (PMMA) was added to aluminum/poly(vinylidene fluoride) (Al/PVDF) energetic blends to enhance melt flow rate and adhesion in a fused deposition modeling (FDM) manufacturing scenario. Composites were successfully printed with up to 30 wt% nano-scale Al after PMMA addition. Melt flow rate increased with increasing PMMA content. This resulted in a partially fluorinated binder that can facilitate high solids loadings and can be easily printed with a standard FDM 3D printer. PMMA addition promoted nucleation of the electroactive β-phase PVDF, which suggests the potential to print piezoelectric, energetic composites. Thermal stability was assessed using differential scanning calorimetry and thermogravimetric analysis. Results verified that composite stability decreased with increasing PMMA and Al content, however, decomposition onset temperatures for all concentrations remained well above printing temperatures. Burn rates at higher Al loadings (e.g., fuel rich) showed a decreasing trend. Analysis of post burn soot revealed α-AlF3 and amorphous carbon char as the primary reaction products. Combustion performance results indicate that although PMMA may serve as a heat sink, the reaction between Al and PVDF was not significantly affected by PMMA addition. These findings indicate that by changing PMMA concentrations, rheology, piezoelectric content, thermal properties and combustion performance can be altered to suit specific needs.
TL;DR: It is illustrated how an integration of aerobic and anaerobic metabolism is required for physiological hypoxia adaptation in skeletal muscle, and protein catabolism and allosteric regulation are highlighted as unexpected orchestrators of metabolic remodeling in this context.
TL;DR: The findings highlight an important area for improving care: routine, documented lethal means assessment and counseling for patients with suicide risk, and an urgent need for further exploration of barriers and facilitators.
Abstract: Prior work from surveys and limited populations suggests many emergency department (ED) patients with suicide risk do not have documented lethal means assessments (e.g., being asked about home firearms). The specific objectives of this study were to, in an ED with universal screening for suicide risk: (1) estimate how often ED providers documented lethal means assessment for suicidal patients, and (2) compare patients with and without documented lethal means assessments. We reviewed 800 total charts from a random sample of adults in three a priori age groups (18–34 years; 35–59 years; ≥ 60 years) with a positive suicide risk screen from 8/2014 to 12/2015. Only 18% (n = 145) had documentation by ≥ 1 provider of assessment of lethal means access. Among these 145, only 8% (n = 11) had documentation that someone discussed an action plan to reduce access (most commonly changing home storage or moving objects out of the home). Among 545 suicidal patients discharged home from the ED, 85% had no documentation that any provider assessed access to lethal means. Our findings highlight an important area for improving care: routine, documented lethal means assessment and counseling for patients with suicide risk. There is an urgent need for further exploration of barriers and facilitators.
TL;DR: In this paper, the inverse scattering transform (IST) with non-zero boundary conditions at infinity is developed for an m × m matrix nonlinear Schrodinger-type equation which, in the case m = 2, has been proposed as a model to describe hyperfine spin F = 1 spinor Bose-Einstein condensates with either repulsive interatomic interactions and anti-ferromagnetic spin-exchange interactions (self-defocusing case), or attractive interatomic interaction and ferromagnetic spins exchange interactions(self-focusing case).
Posted Content•
TL;DR: In this paper, Entropic Open-Set and Objectosphere losses are proposed to maximize entropy for unknown inputs while increasing separation in deep feature space by modifying magnitudes of known and unknown samples.
Abstract: Agnostophobia, the fear of the unknown, can be experienced by deep learning engineers while applying their networks to real-world applications. Unfortunately, network behavior is not well defined for inputs far from a networks training set. In an uncontrolled environment, networks face many instances that are not of interest to them and have to be rejected in order to avoid a false positive. This problem has previously been tackled by researchers by either a) thresholding softmax, which by construction cannot return "none of the known classes", or b) using an additional background or garbage class. In this paper, we show that both of these approaches help, but are generally insufficient when previously unseen classes are encountered. We also introduce a new evaluation metric that focuses on comparing the performance of multiple approaches in scenarios where such unseen classes or unknowns are encountered. Our major contributions are simple yet effective Entropic Open-Set and Objectosphere losses that train networks using negative samples from some classes. These novel losses are designed to maximize entropy for unknown inputs while increasing separation in deep feature space by modifying magnitudes of known and unknown samples. Experiments on networks trained to classify classes from MNIST and CIFAR-10 show that our novel loss functions are significantly better at dealing with unknown inputs from datasets such as Devanagari, NotMNIST, CIFAR-100, and SVHN.
TL;DR: The bystander education program was more effective at changing attitudes, beliefs, efficacy, intentions, and self-reported behaviors compared with the traditional awareness education program, and both programs were significantly more effective than no education.
Abstract: Dating violence is a serious and prevalent public health problem that is associated with numerous negative physical and psychological health outcomes, and yet there has been limited evaluation of prevention programs on college campuses. A recent innovation in campus prevention focuses on mobilizing bystanders to take action. To date, bystander programs have mainly been compared with no treatment control groups raising questions about what value is added to dating violence prevention by focusing on bystanders. This study compared a single 90-min bystander education program for dating violence prevention with a traditional awareness education program, as well as with a no education control group. Using a quasi-experimental pre-test/post-test design with follow-up at 2 months, a sample of predominately freshmen college students was randomized to either the bystander ( n = 369) or traditional awareness ( n = 376) dating violence education program. A non-randomized control group of freshmen students who did not receive any education were also surveyed ( n = 224). Students completed measures of attitudes, including rape myth acceptance, bystander efficacy, and intent to help as well as behavioral measures related to bystander action and victimization. Results showed that the bystander education program was more effective at changing attitudes, beliefs, efficacy, intentions, and self-reported behaviors compared with the traditional awareness education program. Both programs were significantly more effective than no education. The findings of this study have important implications for future dating violence prevention educational programming, emphasizing the value of bystander education programs for primary dating violence prevention among college students.
TL;DR: The second contribution from the OCCAM survey presents analysis of 259 member stars with [Fe/H] determinations in 19 open clusters, using Sloan Digital Sky Survey Data Release 14 (SDSS/DR14) data from the Apache Point Observatory Galactic Evolution Experiment (APOGEE) and ESA Gaia.
Abstract: The Open Cluster Chemical Abundances and Mapping (OCCAM) survey aims to produce a comprehensive, uniform, infrared-based spectroscopic dataset for hundreds of open clusters, and to constrain key Galactic dynamical and chemical parameters from this sample. This second contribution from the OCCAM survey presents analysis of 259 member stars with [Fe/H] determinations in 19 open clusters, using Sloan Digital Sky Survey Data Release 14 (SDSS/DR14) data from the Apache Point Observatory Galactic Evolution Experiment (APOGEE) and ESA Gaia. This analysis, which includes clusters with R_{GC} ranging from 7 to 13 kpc, measures an [Fe/H] gradient of -0.061 \pm 0.004 dex/kpc. We also confirm evidence of a significant positive gradient in the \alpha-elements ([O/Fe], [Mg/Fe], and [Si/Fe]) and present evidence for a significant negative gradient in iron-peak elements ([Mn/Fe] and [Ni/Fe]).