Search

Purpose: This study examines the effectiveness of automatic speech recognition (ASR) for individuals with speech disorders, addressing the gap in performance between read and conversational ASR. We analyze the factors influencing this disparity and the effect of speech mode-specific training on ASR accuracy. Method: Recordings of read and conversational speech from 27 individuals with various speech disorders were analyzed using both (a) one speaker-independent ASR system trained and optimized for typical speech and (b) multiple ASR models that were personalized to the speech of the participants with disordered speech. Word error rates were calculated for each speech model, read versus conversational, and subject. Linear mixed-effects models were used to assess the impact of speech mode and disorder severity on ASR accuracy. We investigated nine variables, classified as technical, linguistic, or speech impairment factors, for their potential influence on the performance gap. Results: We found a significant performance gap between read and conversational speech in both personalized and unadapted ASR models. Speech impairment severity notably impacted recognition accuracy in unadapted models for both speech modes and in personalized models for read speech. Linguistic attributes of utterances were the most influential on accuracy, though atypical speech characteristics also played a role. Including conversational speech samples in model training notably improved recognition accuracy. Conclusions: We observed a significant performance gap in ASR accuracy between read and conversational speech for individuals with speech disorders. This gap was largely due to the linguistic complexity and unique characteristics of speech disorders in conversational speech. Training personalized ASR models using conversational speech significantly improved recognition accuracy, demonstrating the importance of domain-specific training and highlighting the need for further research into ASR systems capable of handling disordered conversational speech effectively. [ABSTRACT FROM AUTHOR]

Background: The environment shapes health behaviors and outcomes. Studies exploring this influence have been limited to research groups with the geographic information systems expertise required to develop built and social environment measures (eg, groups that include a researcher with geographic information system expertise)., Objective: The goal of this study was to develop an open-source, user-friendly, and privacy-preserving tool for conveniently linking built, social, and natural environmental variables to study participant addresses., Methods: We built the automatic context measurement tool (ACMT). The ACMT comprises two components: (1) a geocoder, which identifies a latitude and longitude given an address (currently limited to the United States), and (2) a context measure assembler, which computes measures from publicly available data sources linked to a latitude and longitude. ACMT users access both of these components using an RStudio/RShiny-based web interface that is hosted within a Docker container, which runs on a local computer and keeps user data stored in local to protect sensitive data. We illustrate ACMT with 2 use cases: one comparing population density patterns within several major US cities, and one identifying correlates of cannabis licensure status in Washington State., Results: In the population density analysis, we created a line plot showing the population density (x-axis) in relation to distance from the center of the city (y-axis, using city hall location as a proxy) for Seattle, Los Angeles, Chicago, New York City, Nashville, Houston, and Boston with the distances being 1000, 2000, 3000, 4000, and 5000 m. We found the population density tended to decrease as distance from city hall increased except for Nashville and Houston, 2 cities that are notably more sprawling than the others. New York City had a significantly higher population density than the others. We also observed that Los Angeles and Seattle had similarly low population densities within up to 2500 m of City Hall. In the cannabis licensure status analysis, we gathered neighborhood measures such as age, sex, commute time, and education. We found the strongest predictive characteristic of cannabis license approval to be the count of female children aged 5 to 9 years and the proportion of females aged 62 to 64 years who were not in the labor force. However, after accounting for Bonferroni error correction, none of the measures were significantly associated with cannabis retail license approval status., Conclusions: The ACMT can be used to compile environmental measures to study the influence of environmental context on population health. The portable and flexible nature of ACMT makes it optimal for neighborhood study research seeking to attribute environmental data to specific locations within the United States., (©Weipeng Zhou, Amy Youngbloom, Xinyang Ren, Brian E Saelens, Sean D Mooney, Stephen J Mooney. Originally published in JMIR Formative Research (https://formative.jmir.org), 04.10.2024.)

Objectives: To analyze relationships between Medicaid automatic enrollment for child Supplemental Security Income (SSI) recipients and health insurance coverage during transitions., Data Sources and Study Setting: Medical Expenditure Panel Study, 2000-2020 and National Survey for Children with Special Health Care Needs, 2001-2010., Study Design: Leveraging variation in SSI-Medicaid automatic enrollment status across regions and over time, we estimate a regression model to quantify associations between automatic enrollment and insurance coverage. We validate our findings in the NS-CSHCN., Data Collection: Our sample includes children receiving SSI for a disability. We also analyze a subsample of children newly enrolled in SSI., Principal Findings: Automatic enrollment is associated with a statistically significant increase in insurance coverage. Expanding automatic enrollment to all states is associated with increases in Medicaid enrollment of 3% (CI 0.9%-6.7%) among all SSI children and 7% (CI 1.1%-13.9%) among children newly enrolled in SSI. We find similar decreases in uninsurance. Analysis in the NS-CSHCN replicates these findings., Conclusions: Medicaid automatic enrollment policies are associated with increased insurance coverage for SSI children, particularly those transitioning into the program. Medicaid policy defaults could play an important role in reducing administrative burdens to improve children's coverage and access to care., (© 2023 RAND Corporation. Health Services Research published by Wiley Periodicals LLC on behalf of Health Research and Educational Trust.)

Many caregivers wonder when to talk to children about social inequality and racism, often expressing the belief that children do not pay attention to race or inequity. Here, across 5–9-year-old American children (n = 159, Mage = 7.44; 51.6% female, 47.2% male, 1.2% nonconforming or not provided; 59.1% White, 23.3% racial–ethnic minority, 17.6% not provided) and adults (n = 182, 84.1% female, 13.7% male, 2.2% nonconforming or not provided, 54.9% White, 44.5% racial–ethnic minority, 0.60% not provided), the data show that even 5-year-olds automatically encoded (i.e., spontaneously noticed and remembered) information about race, occupational status, and gender. Although children and adults encoded gender and occupational status at similar levels, adults encoded race at significantly higher levels than children. Additionally, occupational status encoding in children was sensitive to relevant environmental input (i.e., children in less affluent communities were more likely to encode occupational status) suggesting that children become "tuned in" and motivated to notice relevant, salient aspects of their environment. Given the early emergence of gender, race, and occupational status encoding, caregivers should assume that children do notice racial, occupational, or gender inequities in their environments and feel confident in the decision to begin discussing the systemic roots of inequity with children from a young age. Public Significance Statement: This study suggests that children as young as 5 years old spontaneously encode (i.e., automatically notice and remember) race, gender, and occupation—a potential marker of social status. Children from less affluent communities were more likely to encode occupational status than their peers from more affluent communities, suggesting that encoding becomes "tuned in" to relevant aspects of the local context. [ABSTRACT FROM AUTHOR]

The reliable identification of skin and soft tissue infections (SSTIs) from electronic health records is important for a number of applications, including quality improvement, clinical guideline construction, and epidemiological analysis. However, in the United States, types of SSTIs (e.g. is the infection purulent or non-purulent?) are not captured reliably in structured clinical data. With this work, we trained and evaluated a rule-based clinical natural language processing system using 6,576 manually annotated clinical notes derived from the United States Veterans Health Administration (VA) with the goal of automatically extracting and classifying SSTI subtypes from clinical notes. The trained system achieved mention- and document-level performance metrics of the range 0.39 to 0.80 for mention level classification and 0.49 to 0.98 for document level classification.

Biomedical ontologies are a key component in many systems for the analysis of textual clinical data. They are employed to organize information about a certain domain relying on a hierarchy of different classes. Each class maps a concept to items in a terminology developed by domain experts. These mappings are then leveraged to organize the information extracted by Natural Language Processing (NLP) models to build knowledge graphs for inferences. The creation of these associations, however, requires extensive manual review. In this paper, we present an automated approach and repeatable framework to learn a mapping between ontology classes and terminology terms derived from vocabularies in the Unified Medical Language System (UMLS) metathesaurus. According to our evaluation, the proposed system achieves a performance close to humans and provides a substantial improvement over existing systems developed by the National Library of Medicine to assist researchers through this process., (©2023 AMIA - All rights reserved.)

Introduction: This study compared the Salzmann Index (SI) to the American Association of Orthodontists-proposed Automatic Qualifying Conditions (AQC) list in determining eligibility for medically necessary orthodontic care (MNOC). In Pennsylvania, The SI must reflect a score of ≥25, whereas the AQCs are considered binary indicators of a treatment need-the presence of any AQC qualifies a patient for treatment under Medicaid. This study was designed to determine whether the same patients are selected as eligible for MNOC regardless of which selection method is used., Methods: The orthodontic records of 139 participants aged 10-17 years were used to calculate SI scores and identify any AQCs present. McNemar's test of agreement was used to compare patients selected for treatment needs on the basis of SI to those selected on the basis of the presence of an AQC., Results: The mean SI score was 14.7. Twelve patients (8.6%) were selected for treatment needs because of an SI of ≥25, whereas 44 patients (31.7%) were deemed to have at least 1 AQC present. Of the 44 patients with AQCs, only 11 had SI scores of ≥25. McNemar's test of agreement found that SI score and AQC presence did not select patients similarly (P <0.001)., Conclusions: The SI and AQCs do not appear to select the same patients for MNOC, suggesting that replacing the SI with a list of AQCs would significantly impact which patients are selected for Medicaid-funded orthodontic care in Pennsylvania., (Copyright © 2023 American Association of Orthodontists. Published by Elsevier Inc. All rights reserved.)

The vulnerability of Positioning, Navigation, and Timing (PNT) services derived from Global Navigation Satellite Systems (GNSS) makes having a resilient and accurate Alternative PNT (APNT) based on high‐power terrestrial radio sources necessary. The L‐band is very crowded spectral real estate with GNSS, Distance Measuring Equipment (DME) and Air Traffic Control Beacon System (ATCRBS) signals occupying the band from 900–1600 MHz. Thus, as getting new signal and spectrum for APNT would be difficult, we must leverage existing transmissions and infrastructure. The ∼660 Automatic Dependent Surveillance‐Broadcast (ADS‐B) ground stations in the United States represent significant infrastructure that can be leveraged for APNT. However, as ADS‐B was designed for surveillance, it does not inherently possess features necessary to support APNT goals. This paper describes and demonstrates techniques for using ADS‐B Universal Access Transceiver (UAT) signals for PNT. We develop methods to use all ground UAT signals to provide robust, multi‐frequency pseudoranges. We examine the ranging and positioning performance of the UAT signal on the ground and in flight, to demonstrate its ranging accuracy, and hence the timing and synchronization of the station. We demonstrate and analyze navigation using UAT signals, as well as the intra‐system interference challenges of using multiple UAT stations. [ABSTRACT FROM AUTHOR]

Purpose: Point-of-care ultrasound (POCUS) is a sensitive, safe, and efficient tool used in many clinical settings and is an essential part of medical education in the United States. Numerous studies present improved diagnostic performances and positive clinical outcomes among POCUS users. However, others stress the degree to which the modality is user-dependent, rendering high-quality POCUS training necessary in medical education. In this study, the authors aimed to investigate the potential of an artificial intelligence (AI) based quality indicator tool as a teaching device for cardiac POCUS performance. Methods: The authors integrated the quality indicator tool into the pre-clinical cardiac ultrasound course for 4th-year medical students and analyzed their performances. The analysis included 60 students who were assigned to one of two groups as follows: the intervention group using the AI-based quality indicator tool and the control group. Quality indicator users utilized the tool during both the course and the final test. At the end of the course, the authors tested the standard echocardiographic views, and an experienced clinician blindly graded the recorded clips. Results were analyzed and compared between the groups. Results: The results showed an advantage in quality indictor users' median overall scores (P = 0.002) with a relative risk of 2.3 (95% CI: 1.10, 4.93, P = 0.03) for obtaining correct cardiac views. In addition, quality indicator users also had a statistically significant advantage in the overall image quality in various cardiac views. Conclusions: The AI-based quality indicator improved cardiac ultrasound performances among medical students who were trained with it compared to the control group, even in cardiac views in which the indicator was inactive. Performance scores, as well as image quality, were better in the AI-based group. Such tools can potentially enhance ultrasound training, warranting the expansion of the application to more views and prompting further studies on long-term learning effects. [ABSTRACT FROM AUTHOR]

In recent years, a number of states have sought to close the retirement savings funding gap by enacting legislation mandating that employers that do not sponsor a voluntary pension plan for their employees automatically enroll their employees in a stateadministered IRA program. This Article focuses on the most serious legal challenge these programs face: ERISA preemption. The Article begins by providing an overview of the state automatic enrollment IRA programs. It then discusses a regulatory safe harbor created for these programs in 2016 and disapproved under the Congressional Review Act in 2018. It then turns to the question whether, in the absence of the safe harbor, state automatic enrollment IRA s are preempted by ERISA. In addressing this question, it considers two subsidiary issues: (1) whether state automatic enrollment IRA programs are employee benefit plans for purposes of ERISA and thus preempted by ERISA, and (2) even if state automatic enrollment IRA programs are not employee benefit plans for purposes of ERISA, whether the state law creating these programs nevertheless relates to other employee benefit plans so as to be preempted by ERISA. Finally, the Article considers the merits of two complaints that have been filed challenging state automatic enrollment IRA programs. [ABSTRACT FROM AUTHOR]

Introduction: To compare the morphological classification ability of peripheral‐blood leukocytes of the automatic cell morphology analyzers MC‐100i and DI‐60. Methods: (1) MC‐100i and DI‐60 were used to analyze leukocytes in 432 venous blood samples collected from three tertiary hospitals across China. The preclassification results were compared with the results reported by senior morphological experts (postclassification results) to evaluate the accuracy, sensitivity, specificity, and consistency of leukocyte preclassification for both instruments. (2) In 200 of the 432 blood samples, morphological experts conducted manual microscopic examination for various types of leukocytes. The correlation between the MC‐100i and DI‐60 leukocyte postclassification results and the expert microscopist results were analyzed. Results: (1) MC‐100i preclassified leukocytes and nucleated red blood cells (RBCs). Compared with the postclassification results, the total leukocyte preclassification accuracy of MC‐100i was 97.16%, while that of DI‐60 was 87.24%. The sensitivity of MC‐100i to abnormal cells (including blasts, promyelocytes, neutrophilic myelocytes, neutrophilic metamyelocytes, reactive lymphocytes, abnormal promyelocytes, plasma cells, abnormal lymphocytes and nucleated RBCs) was 90.24%, which was significantly higher than the 50.72% sensitivity of DI‐60. (2) Comparing the postclassification results with manual microscopy, except for reactive lymphocytes and basophils, the MC‐100i and DI‐60 results had good correlations with various leukocyte types and nucleated RBCs (r > 0.85), and MC‐100i was better than DI‐60 in the recognition of basophils. Conclusion: Both MC‐100i and DI‐60 have good detection ability for five normal types of leukocytes in peripheral blood. MC‐100i has significantly better detection sensitivity for abnormal cells in peripheral blood than DI‐60. [ABSTRACT FROM AUTHOR]

Voice assistants (VAs), such as Alexa, Siri, or Google Now, are becoming increasingly ubiquitous. Consequently, the potential societal impacts of such systems are gaining relevance in public and academic discourses. Investigating the effects of VA voice features on user perception of VAs is one central aspect of research. Here, the potential effects of VAs on the perception of gender roles and the attribution of gender stereotypes stand out. Following the Computers as Social Actors paradigm, it is assumed that gendered voices have the potential to reproduce existing gender stereotypes. Yet, recently, the implementation of gender-ambiguous voices has been suggested to mitigate such effects and promote more diverse technology design. In this study, we set out to inquire about the relationship between gendered VA voice (male, female, gender-ambiguous) and gender stereotype activation. In two online surveys with an experimental manipulation (Study 1: N = 140, Study 2: N = 397), we test stereotype activation as a function of VA voice, user gender, and conversation topic. Empirical findings are mixed, with no effects for conversation topic but indications of small effects of VA voice. Implications for future research are discussed. [ABSTRACT FROM AUTHOR]

Purpose: The Speech Accessibility Project (SAP) intends to facilitate research and development in automatic speech recognition (ASR) and other machine learning tasks for people with speech disabilities. The purpose of this article is to introduce this project as a resource for researchers, including baseline analysis of the first released data package. Method: The project aims to facilitate ASR research by collecting, curating, and distributing transcribed U.S. English speech from people with speech and/or language disabilities. Participants record speech from their place of residence by connecting their personal computer, cell phone, and assistive devices, if needed, to the SAP web portal. All samples are manually transcribed, and 30 per participant are annotated using differential diagnostic pattern dimensions. For purposes of ASR experiments, the participants have been randomly assigned to a training set, a development set for controlled testing of a trained ASR, and a test set to evaluate ASR error rate. Results: The SAP 2023-10-05 Data Package contains the speech of 211 people with dysarthria as a correlate of Parkinson's disease, and the associated test set contains 42 additional speakers. A baseline ASR, with a word error rate of 3.4% for typical speakers, transcribes test speech with a word error rate of 36.3%. Fine-tuning reduces the word error rate to 23.7%. Conclusions: Preliminary findings suggest that a large corpus of dysarthric and dysphonic speech has the potential to significantly improve speech technology for people with disabilities. By providing these data to researchers, the SAP intends to significantly accelerate research into accessible speech technology. [ABSTRACT FROM AUTHOR]

During the 2022 open enrollment period in California's Affordable Care Act Marketplace, we tested two interventions designed to reduce choice errors among low-income households enrolled in bronze plans that were eligible for zero-premium cost-sharing reduction (CSR) silver plans with more generous benefits. A randomized controlled trial nudge intervention used letter and email reminders to encourage consumers to switch plans, and a quasi-experimental crosswalk intervention automatically enrolled eligible households from bronze plans into zero-premium CSR silver plans with the same insurers and provider networks. The nudge intervention led to a statistically significant 2.3-percentage-point (26 percent) increase in CSR silver plan take-up relative to the control group, but nearly 90 percent of households remained in nonsilver plans. The automatic crosswalk intervention resulted in an 83.0-percentage-point (822 percent) increase in CSR silver plan take-up compared with the control group, with more than 90 percent of households enrolled in CSR silver plans. Our findings can inform health policy debates on the relative effectiveness of different approaches to reducing choice errors among low-income households in the Affordable Care Act Marketplaces.

Chromatographic data processing has garnered attention due to multiple Food and Drug Administration 483 citations and warning letters, highlighting the need for a robust technological solution. The healthcare industry has the potential to greatly benefit from the adoption of digital technologies, but the process of implementing these technologies can be slow and complex. This article presents a "Digital by Design" managerial approach, adapted from pharmaceutical quality by design principles, for designing and implementing an artificial intelligence (AI)-based solution for chromatography peak integration process in the healthcare industry. We report the use of a convolutional neural network model to predict analytical variability for integrating chromatography peaks and propose a potential GxP framework for using AI in the healthcare industry that includes elements on data management, model management, and human-in-the-loop processes. The component on analytical variability prediction has a great potential to enable Industry 4.0 objectives on real-time release testing, automated quality control, and continuous manufacturing., (© 2023 Merck Serono S.p.A. Biotechnology and Bioengineering published by Wiley Periodicals LLC.)

Purpose: The present study examined patterns of demographic characteristics that are predictive of people's automatic preference for people without disabilities over people with disabilities. Method: Predictions of automatic preference for people without disabilities were investigated among 41,789 individuals who volunteered to complete the Disability Implicit Association Test (D-IAT) at the Project Implicit website (https://implicit.harvard.edu/). Results: Results indicate that men over 55 without a disability, who scored neutral or high on explicit preference for people without disabilities, who are religious and conservative, are likely to have higher automatic and explicit preference for people without disabilities. Conclusions: Results of the present study provide support for the use of implicit association measures to assess latent attitudes toward people with and without disabilities. [ABSTRACT FROM AUTHOR]

The article discusses the automatic stay legal doctrine which protects debtors during a bankruptcy in America, and it mentions a debate about what, if any, automatic stay remains for a repeat bankruptcy filer under the U.S. Bankruptcy Code after the termination of an automatic stay. America's Bankruptcy Abuse Prevention and Consumer Protection Act of 2005 is examined, along with the views of a majority of courts in the U.S. America's First Circuit Court of Appeals is assessed.

In the United States, many individuals with diabetes mellitus (DM) do not achieve treatment goals despite the availability of effective interventions. Provider clinical inertia is one cause of these unfavorable outcomes. Targeted automatic eConsults (TACos) are an emerging technology-based intervention with potential to address clinical inertia in primary care (PC). TACos prospectively identify at-risk patients and use unsolicited specialist recommendations to prompt treatment intensification. Through a payer-provider collaboration, a TACos intervention was piloted for adults with uncontrolled DM (HbA1c >8%) to understand impact on DM clinical inertia and outcomes. Clinical inertia was assessed by measuring whether a PC provider implemented recommended therapeutic changes. Six-month changes in HbA1c and health care costs per member per month were evaluated using an observational matched design and intention-to-treat (ITT) analysis. The analysis included 196 individuals who received a TACos between February 2021 and August 2021 (ITT group) matched to 392 controls based on clinical and demographic criteria. TACos recommendations were implemented 65% of the time. Median percent change in HbA1c was significantly greater for the ITT group versus controls (-10.9% vs. -10.2%; P  = 0.0359). Median total costs were 7.9% lower in the ITT group ( P  = 0.0900). A per protocol analysis was done to examine effects between ITT group individuals with an implemented TACos recommendation ( n  = 126) and controls. Median percent change in HbA1c was significantly greater (-19.5% vs. -10.2%; P  < 0.0001), but there was no difference in total costs (-7.9%; P  = 0.1753). TACos may feasibly address clinical inertia in PC and improve HbA1c for uncontrolled DM.

Background: Pediatric palliative care (PPC) consultation is infrequent among children on extracorporeal membrane oxygenation (ECMO). Objective: Investigate intensive care unit (ICU) team members' perceptions of automatic PPC consultation for children on ECMO in an ICU in the United States. Methods: Cross-sectional survey assessing benefits, barriers to PPC, and consultation processes. Results: Of 291 eligible respondents, 48% (n = 140) completed the survey and 16% (n = 47) answered an open-ended question. Benefits included support in decision-making (n = 98; 70%) and identification of goals of care (n = 89; 64%). Barriers included perception of giving up on families (n = 59; 42%) and poor acceptability by other team members (n = 58; 41%). Respondents endorsed communication with the primary ICU team before (n = 122; 87%) and after (n = 129; 92%) consultation. Open-ended responses showed more positive (79% vs. 13%) than negative statements. Positive statements reflected on expanding PPC to other critically-ill children where negative statements revealed unrecognized value in PPC. Conclusions: Results demonstrate opportunities for education about the scope of PPC and improvements in PPC delivery. [ABSTRACT FROM AUTHOR]

Daphnia magna is an important organism in ecotoxicity studies because it is sensitive to toxic substances and easy to culture in laboratory conditions. Its locomotory responses as a biomarker are highlighted in many studies. Over the last several years, multiple high-throughput video tracking systems have been developed to measure the locomotory responses of Daphnia magna. These high-throughput systems, used for high-speed analysis of multiple organisms, are essential for efficiently testing ecotoxicity. However, existing systems are lacking in speed and accuracy. Specifically, speed is affected in the biomarker detection stage. This study aimed to develop a faster and better high-throughput video tracking system using machine learning methods. The video tracking system consisted of a constant temperature module, natural pseudo-light, multi-flow cell, and an imaging camera for recording videos. To measure Daphnia magna movements, we developed a tracking algorithm for automatic background subtraction using k-means clustering, Daphnia classification using machine learning methods (random forest and support vector machine), and tracking each Daphnia magna location using the simple online real-time tracking algorithm. The proposed tracking system with random forest performed the best in terms of identification (ID) precision, ID recall, ID F1 measure, and ID switches, with scores of 79.64%, 80.63%, 78.73%, and 16, respectively. Moreover, it was faster than existing tracking systems such as Lolitrack and Ctrax. We conducted an experiment to observe the impact of toxicants on behavioral responses. Toxicity was measured manually in the laboratory and automatically using the high-throughput video tracking system. The median effective concentration of Potassium dichromate measured in the laboratory and using the device was 1.519 and 1.414, respectively. Both measurements conformed to the guideline provided by the Environmental Protection Agency of the United States; therefore, our method can be used for water quality monitoring. Finally, we observed Daphnia magna behavioral responses in different concentrations after 0, 12, 18, and 24 h and found that there was a difference in movement according to the concentration at all hours., (© 2023. The Author(s).)

We analyze how Medicaid automatic enrollment policies for children with special health care needs (CSHCN) who are enrolled in Supplemental Security Income (SSI) reduce disparities in health insurance coverage and caregiving burden. Using the 2009-2010 National Survey of Children with Special Health Care Needs, we implement a difference-in-differences regression model comparing insurance enrollment rates between CSHCN receiving SSI and CSHCN not receiving SSI, in states with and without automatic enrollment policies. We find that Medicaid automatic enrollment has a meaningful impact on insurance enrollment for low-income CSHCN who participate in SSI and can be an effective method for mitigating disparities in insurance coverage (reducing uninsurance by 38%). Medicaid automatic enrollment also reduces caregiver burden among socioeconomically disadvantaged families with CSHCN. The effects of these policies are largest families who might be on the margin of eligibility or who face high administrative burden.

Product-specific guidances (PSGs) recommended by the United States Food and Drug Administration (FDA) are instrumental to promote and guide generic drug product development. To assess a PSG, the FDA assessor needs to take extensive time and effort to manually retrieve supportive drug information of absorption, distribution, metabolism, and excretion (ADME) from the reference listed drug labeling. In this work, we leveraged the state-of-the-art pre-trained language models to automatically label the ADME paragraphs in the pharmacokinetics section from the FDA-approved drug labeling to facilitate PSG assessment. We applied a transfer learning approach by fine-tuning the pre-trained Bidirectional Encoder Representations from Transformers (BERT) model to develop a novel application of ADME semantic labeling, which can automatically retrieve ADME paragraphs from drug labeling instead of manual work. We demonstrate that fine-tuning the pre-trained BERT model can outperform conventional machine learning techniques, achieving up to 12.5% absolute F1 improvement. To our knowledge, we were the first to successfully apply BERT to solve the ADME semantic labeling task. We further assessed the relative contribution of pre-training and fine-tuning to the overall performance of the BERT model in the ADME semantic labeling task using a series of analysis methods, such as attention similarity and layer-based ablations. Our analysis revealed that the information learned via fine-tuning is focused on task-specific knowledge in the top layers of the BERT, whereas the benefit from the pre-trained BERT model is from the bottom layers., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2023 Elsevier Inc. All rights reserved.)

Background: Telemedicine support virtual consultations and evaluations in hand surgery for patients in remote areas during the COVID-19 era. However, traditional physical examination is challenging in telemedicine and it is inconvenient to manually measure the hand range of motion (ROM) from images or videos. Here, we propose an automatic method using the hand pose estimation technique, aiming to measure the hand ROM from smartphone images., Methods: Twenty-eight healthy volunteers participated in the study. An eight-hand gestures measurement protocol and the Google MediaPipe Hands were used to analyze images and calculate the ROM automatically. Manual goniometry was also performed according to the guideline of the American Medical Association. The correlation between the automatic and manual methods was analyzed by the intraclass correlation coefficient and Pearson correlation coefficient. The clinical acceptance was testified using Bland-Altman plots., Results: A total of 32 parameters of each hand were measured by both methods, and 1792 measurement results were compared. The mean difference between automatic and manual methods is -2.21 ± 9.29° in the angle measurement and 0.48 ± 0.48 cm in the distance measurement. The intraclass correlation coefficient of 75% of parameters was higher than 0.75, the Pearson correlation coefficient of 84% of parameters was over 0.6, and 40.6% of parameters reached well-accepted clinical agreements., Conclusions: The proposed method provides a helpful protocol for automatic hand ROM measurement based on smartphone images and the MediaPipe Hands pose estimation technique. The automatic measurement is acceptable and comparable with existing methods, showing a possible application in the telemedicine examination of hand surgery.

Automatic item generation represents a relatively new but rapidly evolving research area where cognitive and psychometric theories are used to produce tests that include items generated using computer technology. Automatic item generation requires two steps. First, test development specialists create item models, which are comparable to templates or prototypes, that highlight the features or elements in the assessment task that must be manipulated. Second, these item model elements are manipulated to generate new items with the aid of computer-based algorithms. With this two-step process, hundreds or even thousands of new items can be created from a single item model. The purpose of our article is to describe seven different but related topics that are central to the development and use of item models for automatic item generation. We start by defining item model and highlighting some related concepts; we describe how item models are developed; we present an item model taxonomy; we illustrate how item models can be used for automatic item generation; we outline some benefits of using item models; we introduce the idea of an item model bank; and finally, we demonstrate how statistical procedures can be used to estimate the parameters of the generated items without the need for extensive field or pilot testing. (Contains 4 notes, 2 tables and 4 figures.)

Background and Aims: Prescription drug monitoring programs (PDMPs) were implemented to decrease dangerous opioid prescribing but have had variable results. This report details how automatic PDMP review changed opioid prescribing across a statewide medical system., Design: An observational study., Setting: Fourteen hospital networks in the United States., Cases: Healthcare encountered from July 1, 2016 to June 30, 2019., Intervention: Starting from July 1, 2018, the patient's PDMP data would be displayed automatically to providers in the unified electronic medical record (EMR) whenever the provider began to write for an opioid prescription., Measurements: Outcomes were prescriptions per encounter (PPE) and the morphine milligram equivalents (MME) per prescription. Outcomes were stratified by practice location, medication prescribed, and diagnosis. All data, including whether the prompt was triggered for a given encounter and whether a prescription was given, were extracted from the EMR. An interrupted timeseries analysis was used to determine how PPE and MME changed in response to the implementation of automatic PDMP review., Findings: Of the 624,488 encounters examined, 18.37 percent (n = 114,710) were in emergency departments, 56.79 percent were admissions (n = 354,634), and 24.84 percent (n = 155,144) were outpatient visits. Opioid prescriptions were started and then canceled 24 percent of the time after the PDMP was shown. There was a decline in MME (βOverall + Policy Trends = -3.17, p = <0.0001), which was driven by inpatient (βOverall + Policy Trends = -2.10, p < 0.0001) and outpatient providers (βOverall + Policy Trends = -3.24, p < 0.01). A decline in MME was seen in all medication categories (-1.72 < βOverall + Policy Trends < -5.87, p < 0.01). There were no changes in these trends after excluding encounters for severe and acute pain., Conclusions: Automated PDMP review is associated with fewer prescriptions and smaller doses without decreasing appropriate use.

Many cotton breeding programs in the U.S. have installed an automatic weighing system in a cotton plot picker to increase operating efficiency in recent years. However, no experimental data are available to document the reliability of such a system. The objective of this study was to evaluate the accuracy, precision, and harvesting efficiency of a two-row cotton plot picker installed with an automatic weighing system with two scales based on results from replicated field tests from 2013 to 2016. Three tests each year, each containing 32 genotypes were arranged in a randomized complete block design in two-row plots by 10 m in length. The 2013 to 2015 results showed a highly significant positive correlation in seedcotton weights between the two rows of the same plot for each genotype harvested and weighed by the two scales in the picker, indicating that the two scales are consistent and reliable. In three tests in 2016, one row of each two-row plot was harvested by the picker, and seedcotton weight was compared with another row harvested by hand. A highly significant positive correlation was detected between the two harvesting methods that had similar coefficients of variation (16.14% for hand harvest vs. 16.90% for mechanical harvest). The two-row plot picker harvested a total of 368 single-row plots (10-m long) in six hours daily, whereas hand harvest by one person averaged two plots in four hours. An average of 417 kg ha-1 was lost due to the mechanical harvest. [ABSTRACT FROM AUTHOR]

Background: Contracts can be an effective lever to implement and manage health-enabling food retail environments. However, guidance for the effective use of contracts in food retail settings is limited. The use of contracts to create healthy food vending environments is one area where policy attention has been focussed in high income countries. We applied a public health regulatory framework to publicly available guidance documents on healthy vending to develop best practice recommendations for using contracts to create healthy food vending environments., Methods: Document analysis involved i) snowball sampling to identify eligible publicly available healthy vending guidance documents from an identified seed paper; ii) application of a public health regulatory framework to extract data across three domains of form, substance and governance of healthy vending initiatives; and iii) synthesis of data to form best practice recommendations. Eligible documents were those aimed at implementing healthier vending; published from 2000 onwards; accessible online; and included recommendations beyond nutrition standards alone, including a reference to at least one regulatory governance process (administration, implementation, monitoring, enforcement or review)., Results: Twelve of 92 documents identified were eligible and all were from the United States (US). All noted that products need to comply with nutrition standards. Other aspects of regulatory substance (i.e., pricing, promotion, placement, labelling and contract length) were less well considered as were elements of regulatory governance (regulatory rules, administration, implementation, monitoring, enforcement and review). Our adapted framework covers three regulatory domains with nine components, and a further 20 recommendations for best practice application in healthy vending., Conclusions: To be effective, contracts used to manage healthy food vending should include more than the nutrition standards for healthy food and drinks. Clearly stating the contract objectives, operative terms and conditions, and defining responsibilities for monitoring, review and enforcement within the contract, in addition to the nutrition standards, will assist practitioners in creating effective and sustained contract-based initiatives aimed at improving the healthiness of vending, or potentially other food retail environments., Competing Interests: Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests., (© 2024. The Author(s).)

Objective: Forward collision warning and automatic emergency braking (AEB) systems help prevent rear-end crashes where a vehicle strikes the rear of another. However, the benefits may be limited if the systems are stymied by common crash circumstances or only target scenarios evaluated in vehicle testing programs. This study examined the prevalence of characteristics that may limit AEB system performance in police-reported rear-end crashes and the relevance of scenarios used to evaluate these systems., Methods: Police-reported rear-end crashes ( n  = 6,731,215, Crash Report Sampling System) and fatal rear-end crashes ( n  = 4,285, Fatality Analysis Reporting System) with a fatality in the striking or struck vehicle during 2016-2019 were analyzed. Percentages of police-reported rear-end crashes, nonfatal-injury rear-end crashes, and fatal rear-end crashes were computed to identify common crash characteristics. Roadway speed limit was used as a proxy for striking vehicle speed., Results: A straight-moving vehicle striking a stopped or decelerating vehicle on roads with a speed limit of 40 km/h (25 mph) or less only accounted for 3% of all rear-end crashes, 3% of nonfatal-injury rear-end crashes, and 1% of fatal rear-end crashes. In contrast, 36% of all rear-end crashes, 36% of nonfatal-injury rear-end crashes, and 11% of fatal rear-end crashes involved a straight-moving vehicle striking a stopped or decelerating vehicle on roads with a speed limit between 56 and 72 km/h (35 and 45 mph). A medium or heavy truck was the struck vehicle in 32% of fatal rear-end crashes, and a motorcycle was the struck vehicle in 11% of fatal rear-end crashes. At least one of the following characteristics that may degrade AEB system performance was present in 14% of the rear-end crashes studied: striking vehicle turning; a struck vehicle turning or changing lanes; a struck vehicle that is not a passenger vehicle; wintery weather; wet or icy roads; or a speed limit of 113 km/h (70 mph) or higher., Conclusion: Circumstances shown to diminish AEB effectiveness accounted for 14% of rear-end crashes, and scenarios currently used to evaluate AEB systems accounted for 3%. Evaluating AEB systems at speeds up to 72 km/h (45 mph) and incorporating a motorcycle or medium/heavy truck target will make AEB evaluations more representative of police-reported rear-end crashes.

Resilience, a psychological trait conceptualized as the ability to recover from setbacks, can be weakened by childhood maltreatment, which comprises childhood abuse and childhood neglect. The current study aimed to investigate whether childhood maltreatment could increase automatic negative thoughts (ANT), thus weakening resilience. Furthermore, as psychological characteristics are commonly subject to the moderating effects of cultural context and biology, the study also explored whether and how cultural and genetic factors separately interact with childhood maltreatment to predict resilience. In study 1, the participants comprised 237 American and 347 Chinese individuals; study 2 included 428 genotyped Chinese individuals. We combined regression, mediation, moderation, and machine learning methods to test the mediating effect of ANT on the link between childhood maltreatment and resilience as well as the moderating roles of culture and genetics. Study 1 found that both childhood abuse and childhood neglect increased ANT and thus weakened resilience. In addition, the ANT-mediating effects of childhood neglect were stronger in American than Chinese participants. In Study 2, using the leave-one-out approach, we constructed two separate prediction models based on 22 and 16 important single nucleotide polymorphisms (SNPs), and we found that the interaction between childhood abuse/neglect and polygenic scores based on important SNPs could predict ANT. The mediating effects of ANT on the relationship between childhood abuse/neglect and resilience were significant for participants with low polygenic scores but not for those with high polygenic scores. In conclusion, both the cultural environment and individual genetic makeup moderated the mediating effects of ANT on the association between childhood maltreatment and resilience. These findings indicated the roles of culture and genetics in protecting against the adverse effects of childhood maltreatment on resilience.

Understanding how field geometry is related to over-application of inputs may help farmers make more informed decisions about investment in automatic section control (ASC) technology. The technology helps reduce input application overlap and therefore reduces input costs. The influence of field geometry on the profitability of ASC technology for chemical application was evaluated using overlap data estimated for 44 farm fields in Tennessee, USA. Reduction in overlap with ASC was defined as the difference in area sprayed controlling the entire boom as one section and area sprayed using seven-section ASC. Perimeter (m)-to-area (m) ratio (P/A, m) was used to categorize estimated chemical overlap by field size and shape. Estimated median reductions in overlap with ASC were 3.00 % for fields with P/A = 0.01, 9.65 % for fields with P/A = 0.02 and 13.50 % for fields with P/A ≥ 0.03. For a typical size cotton farm in Tennessee, investing in ASC was not likely to be profitable for fields with P/A = 0.01 but was generally profitable for fields with P/A ≥ 0.02. The low reduction in overlap for P/A = 0.01 with ASC resulted in chemical savings that were too small to pay back the investment in ASC within the useful life assumed for the technology and provide a minimum rate of return. Thus, P/A as a measure of field geometry may be useful for evaluating investments in ASC technologies. [ABSTRACT FROM AUTHOR]

Within three years of the invention of the telephone in 1876 attempts were being made for the partial machine switching of calls. A step-by-step system was developed by A.B. Strowger of the USA in the years 1889 to 1892 and from the beginning of the twentieth century was the first automatic exchange system to come into widespread use there. Conditions in Britain at the time were not conducive to the adoption of machine switching while telephone service was divided between the Post Office and the National Telephone Company (NTC) but the Post Office undertook a review of USA practice. The Post Office completed the purchase of NTC in January 1912 and instituted a series of trials of available automatic systems. The first of these, the public exchange at Epsom, Surrey, opened in May 1912 using the Strowger system. The paper describes the installation and operation of the equipment at Epsom and how it was received by its users and the wider public. In 1922 the Post Office decided to standardize on a development of Strowger step-by-step for British requirements. Ironically, when Epsom exchange required enlargement in the early 1930s it was instead replaced by a new manual exchange. A few examples of equipment from the pioneer British automatic exchange systems survive in preservation. [ABSTRACT FROM AUTHOR]

Objective:In order to save teachers' correcting time, improve the accuracy and efficiency of English composition grading.Methods: This paper briefly introduces the algorithm of deep sentence smoothness and text semantic matching based on graph neural network, and then designs an automatic scoring algorithm for English text. Result: The experimental data was collected from 12,000 essays written by international students in the United States in the Pratt & Whitney Foundation's Automated Student Value Assessment Project (ASAP), and these data were graded through a comparative experiment,Through comparative tests, the automatic scoring algorithm designed in this paper can achieve better scoring results and better handle automatic essay scoring problems. Among all the experimental mean values of evaluation methods, the experimental mean value of the algorithm designed in this paper is 0.790, the smoothness algorithm is 0.768, and the text matching vector is 0.759. The experimental mean values of the other two traditional automatic scoring algorithms are 0.710 and 0.712 respectively, and the results are lower than the algorithm designed in this paper. Conclusion: According to the experimental results, it can be concluded that good feature selection can give good scoring performance to the algorithm and cope with the problem of automatic scoring. At the same time, it also confirms the feasibility of the algorithm designed in this paper, which can be effectively applied in practical English composition scoring. [ABSTRACT FROM AUTHOR]

Diabetic retinopathy (DR), one of the major and long-term microvascular complications of diabetes, is the most common cause of vision loss and blindness in the working population of the world. Even with the management of diabetes, most patients will develop some forms of DR after approximately 20 years. However, DR is a treatable disease throughout the disease progression. To provide appropriate DR management, the USA and European countries have successfully implemented systematic early DR screening programs. At the same time, some computer-aided DR screening systems, which combine advanced DR detection algorithms and telemedicine technology, have also been developed for early-stage DR detection. Some of them have been tested in the DR screening programs. In this paper, we focus on a review of the major automatic DR screening systems which have performed large-scale evaluation rather than give an extensive review of all published DR grading algorithms. We first present the structures of the automatic systems and their supporting algorithms developed by the research groups, as well as the practices of the systems in their screening programs. We further present a more detailed review of the DR lesion detection algorithms in each system and reveal how the DR screening systems successfully practiced in clinical trials or large-scale screening programs. We also review recently new research areas as well as deep learning-based DR screening systems and compare them with the traditional lesion detection-based DR screening systems. The performances of the systems in the trials are summarized by considering the specificity and sensitivity with respect to the scale of testing datasets. At last, we will discuss future challenges. [ABSTRACT FROM AUTHOR]

The article discusses the U.S. Court of Appeals for the Tenth Circuit's ruling in the 2013 case Schoppe v. Commissioner which deals with the application of an automatic stay provision under Chapter 11, Section 362(A)(1) of the U.S. Code to appeals from the nation's Tax Court. A split of opinion among various American circuit courts is addressed, along with U.S. bankruptcy laws, American Internal Revenue Service proceedings, and U.S. legislative history. Appellant John H. Schoppe is mentioned.

In the USA, an interchangeability designation provides biosimilar sponsors with a pathway for achieving what is standard for small-molecule generics: pharmacy-level auto-substitution for an innovator. No other major health authority links interchangeability to automatic substitution, as all require the involvement of the prescriber or patient in a medication change. This editorial considers the clinical impact and practicality of auto-substitution. First, interchangeability is linked to non-medical switching (NMS), the practice of switching treatment in patients with stable disease for non-clinical reasons. NMS may generate negative sentiment in those unwilling or reluctant to switch, which can adversely impact treatment outcomes (i.e., nocebo effect). Indeed, in real-world studies of tumor necrosis factor inhibitors, discontinuation rates have been shown to be higher in patients switched to biosimilars for non-medical reasons than in historical cohorts maintained on innovators. Second, interchangeability may impede pharmacovigilance and traceability, as not all jurisdictions require innovators and biosimilars to have distinct biologic names. Third, an interchangeability designation from the US Food and Drug Administration only permits a biosimilar to be automatically substituted for its innovator, not other biosimilars (if available). Pharmacist education would be needed to avoid off-label, automatic substitution among biosimilars of a single innovator. Last, once granted, an interchangeability designation exists in perpetuity under current US federal law. However, the supply chains of innovators and biosimilars are maintained independently, with no requirement for reconfirmation of biosimilarity or interchangeability. We feel that additional guidance is needed for the auto-substitution of biosimilars and innovators to become a reality.

The automatic stay provision is one of the most important provisions in the Bankruptcy Code. Until recently, however, it has remained unclear if passive retention of property of the bankruptcy estate must be immediately turned over to the debtor under the automatic stay provision. The Supreme Court decided in City of Chicago v. Fulton that passive retention does not violate the automatic stay, saving creditors from the consequences of retaining estate property. The debate about the stay, however, is far from over. Many circuit courts were already concerned about the policy issues deriving from the City of Chicago maintaining possession over debtors' vehicles, rendering debtors unable to get to work or pay off their debts during the bankruptcy plan. Justice Sotomayor remains frustrated about these policy concerns and called on Congress to change the Code and enable debtors to be more successful in their payment plans. This Note suggests Congress make language and cross-reference changes to the Bankruptcy Code to satisfy both textual and policy issues associated with the current Code. [ABSTRACT FROM AUTHOR]

Individuals with a laryngectomy face a host of challenges ranging from restricted vocal communication to significant lifestyle modifications associated with breathing through a stoma. Although there are significant mental and physical health benefits achieved by returning to recreational pursuits that were enjoyed pre-surgery, there can be significant obstacles in doing so. One particular challenge arises during participation in water activities (e.g, fishing, boating, etc.) where accidental submersion poses a significant risk of drowning. This manuscript describes a proof-of-concept device that protects the airway from accidental incursion of water into the airway during unanticipated submersion in water, thereby allowing laryngectomees to return to participation in water activities. The device is designed to be worn comfortably for long periods of time, while not interfering with the common methods of replacement speech that are utilized post-laryngectomy., Competing Interests: The authors have declared that no competing interests exist.

Importance: The American Rescue Plan increases premium subsidies for health insurance marketplace enrollees, potentially leading to situations in which enrollees could switch to other health care plans with lower premiums and less cost sharing (ie, deductibles and copayments). Current policy defaults enrollees to their current health care plan if they automatically renew their coverage, which may cause them to stay in health care plans that, because of the American Rescue Plan, are now dominated in that they have higher premiums and cost sharing than other options., Objective: To estimate the extent to which a smart default policy could reduce US health insurance marketplace enrollees' cost sharing and premiums., Design Setting and Participants: Using 2018 individual enrollment data and 2021 premium data from California's marketplace and the American Rescue Plan premium tax credit subsidy schedule, this economic analysis estimated the characteristics of enrollees' default health care plans if they defaulted into 2021 health care plans under current and smart default policies. The analysis was conducted from March 20 to April 8, 2021., Main Outcomes and Measures: Characteristics of enrollees' default health care plans under current and smart default policies, including net premiums, plan levels, and cost sharing., Results: The analytic sample consisted of 748 087 Covered California enrollees from 2018 (mean [SD] age, 44.80 [13.72] years; 408 410 [54.6%] women). Under current policy with the enhanced subsidies implemented under the American Rescue Plan, 5.8% of sample enrollees would default into dominated health plans. Of these enrollees, 98.0% would have incomes below 250% of the federal poverty level. A smart default policy would lead to a mean $102.47 decrease in monthly premiums (95% CI, $103.84-$101.10), a mean $1960 reduction in individual annual medical deductibles (95% CI, $1991-$1928), and a $49.56 reduction in specialty prescription copays (95% CI, $49.77-$49.34)., Conclusions and Relevance: The findings of this economic analysis suggest that a smart default policy could avoid defaulting lower-income marketplace enrollees to objectively inferior health care insurance plans and may lead to large reductions in lower-income enrollees' deductibles, copayments, and maximum out-of-pocket amounts. Implementation of a smart default policy could enable marketplace administrators to reduce the prevalence of underinsurance among lower-income marketplace enrollees., Competing Interests: Conflict of Interest Disclosures: Dr Drake reported receiving grants from the National Institute for Health Care Management for related work outside the submitted work. No other disclosures were reported., (Copyright 2021 Anderson DM et al. JAMA Health Forum.)

In the calibration of hydrological models, evaluation criteria are explicitly and quantitatively defined as single- or multi-objective functions when utilizing automatic calibration approaches. In most previous studies, there is a general opinion that no single-objective function can represent all important characteristics of even one specific hydrological variable (e.g., streamflow). Thus hydrologists must turn to multi-objective calibration. In this study, we demonstrated that an optimized single-objective function can compromise multi-response modes (i.e., multi-objective functions) of the hydrograph, which is defined as summation of a power function of the absolute error between observed and simulated streamflow with the exponent of power function optimized for specific watersheds. The new objective function was applied to 196 model parameter estimation experiment (MOPEX) watersheds across the eastern United States using the semi-distributed Xinanjiang hydrological model. The optimized exponent value for each watershed was obtained by targeting four popular objective functions focusing on peak flows, low flows, water balance, and flashiness, respectively. Results showed that the optimized single-objective function can achieve a better hydrograph simulation compared to the traditional single-objective function Nash-Sutcliffe efficiency coefficient for most watersheds, and balance high flow part and low flow part of the hydrograph without substantial differences compared to multi-objective calibration. The proposed optimal single-objective function can be practically adopted in the hydrological modeling if the optimal exponent value could be determined a priori according to hydrological/climatic/landscape characteristics in a specific watershed. [ABSTRACT FROM AUTHOR]

Accurate and timely information on the distribution of irrigated croplands is crucial to research on agriculture, water availability, land use, and climate change. While agricultural land use has been well characterized, less attention has been paid specifically to croplands that are irrigated, in part due to the difficulty in mapping and distinguishing irrigation in satellite imagery. In this study, we developed a semi-automatic training approach to rapidly map irrigated croplands across the conterminous United States (CONUS) at 30 m resolution using Google Earth Engine. To resolve the issue of lacking nationwide training data, we generated two intermediate irrigation maps by segmenting Landsat-derived annual maximum greenness and enhanced vegetation index using county-level thresholds calibrated from an existing coarse resolution irrigation map. The resulting intermediate maps were then spatially filtered to provide a training data pool for most areas except for the upper midwestern states where we visually collected samples. We then used random samples extracted from the training pool along with remote sensing-derived features and climate variables to train ecoregion-stratified random forest classifiers for pixel-level classification. For ecoregions with a large training pool, the procedure of sample extraction, classifier training, and classification was conducted 10 times to obtain stable classification results. The resulting 2012 Landsat-based irrigation dataset (LANID) identified 23.3 million hectares of irrigated croplands in CONUS. A quantitative assessment of LANID showed superior accuracy to currently available maps, with a mean Kappa value of 0.88 (0.75–0.99), overall accuracy of 94% (87.5–99%), and producer's and user's accuracy of the irrigation class of 97.3% and 90.5%, respectively, at the aquifer level. Evaluation of feature importance indicated that Landsat-derived features played the primary role in classification in relatively arid regions while climate variables were important in the more humid eastern states. This methodology has the potential to produce annual irrigation maps for CONUS and provide insights into the field-level spatial and temporal aspects of irrigation. [ABSTRACT FROM AUTHOR]

Abstract Owing to its recent advance, machine learning has spawned a large collection of solar forecasting works. In particular, machine learning is currently one of the most popular approaches for hourly solar forecasting. Nevertheless, there is evidently a myth on forecast accuracy—virtually all research papers claim superiority over others. Apparently, the "best" model can only be selected with hindsight, i.e., after empirical evaluation. For any new forecasting project, it is irrational for solar forecasters to bet on a single model from the start. In this article, the hourly forecasting performance of 68 machine learning algorithms is evaluated for 3 sky conditions, 7 locations, and 5 climate zones in the continental United States. To ensure a fair comparison, no hybrid model is considered, and only off-the-shelf implementations of these algorithms are used. Moreover, all models are trained using the automatic tuning algorithm available in the R caret package. It is found that tree-based methods consistently perform well in terms of two-year overall results, however, they rarely stand out during daily evaluation. Although no universal model can be found, some preferred ones for each sky and climate condition are advised. Highlights • Hourly solar forecasting is performed using 68 machine learning models. • Models are evaluated at 7 locations in 5 climate zones for 2 years. • Tree-based methods consistently perform well in terms of 2-year average metrics. • Daily best model cannot be identified, regime-switching approach is advised. [ABSTRACT FROM AUTHOR]

Camera traps provide a low-cost approach to collect data and monitor wildlife across large scales but hand-labeling images at a rate that outpaces accumulation is difficult. Deep learning, a subdiscipline of machine learning and computer science, can address the issue of automatically classifying camera-trap images with a high degree of accuracy. This technique, however, may be less accessible to ecologists or small-scale conservation projects, and has serious limitations. In this study, we trained a simple deep learning model using a dataset of 120,000 images to identify the presence of nilgai Boselaphus tragocamelus, a regionally specific nonnative game animal, in camera-trap images with an overall accuracy of 97%. We trained a second model to identify 20 groups of animals and one group of images without any animals present, labeled as "none," with an accuracy of 89%. Lastly, we tested the multigroup model on images collected of similar species, but in the southwestern United States, resulting in significantly lower precision and recall for each group. This study highlights the potential of deep learning for automating camera-trap image processing workflows, provides a brief overview of image-based deep learning, and discusses the often-understated limitations and methodological considerations in the context of wildlife conservation and species monitoring. [ABSTRACT FROM AUTHOR]

Background: Prescription medication (PM) misuse/abuse has emerged as a national crisis in the United States, and social media has been suggested as a potential resource for performing active monitoring. However, automating a social media-based monitoring system is challenging-requiring advanced natural language processing (NLP) and machine learning methods. In this paper, we describe the development and evaluation of automatic text classification models for detecting self-reports of PM abuse from Twitter.Methods: We experimented with state-of-the-art bi-directional transformer-based language models, which utilize tweet-level representations that enable transfer learning (e.g., BERT, RoBERTa, XLNet, AlBERT, and DistilBERT), proposed fusion-based approaches, and compared the developed models with several traditional machine learning, including deep learning, approaches. Using a public dataset, we evaluated the performances of the classifiers on their abilities to classify the non-majority "abuse/misuse" class.Results: Our proposed fusion-based model performs significantly better than the best traditional model (F1-score [95% CI]: 0.67 [0.64-0.69] vs. 0.45 [0.42-0.48]). We illustrate, via experimentation using varying training set sizes, that the transformer-based models are more stable and require less annotated data compared to the other models. The significant improvements achieved by our best-performing classification model over past approaches makes it suitable for automated continuous monitoring of nonmedical PM use from Twitter.Conclusions: BERT, BERT-like and fusion-based models outperform traditional machine learning and deep learning models, achieving substantial improvements over many years of past research on the topic of prescription medication misuse/abuse classification from social media, which had been shown to be a complex task due to the unique ways in which information about nonmedical use is presented. Several challenges associated with the lack of context and the nature of social media language need to be overcome to further improve BERT and BERT-like models. These experimental driven challenges are represented as potential future research directions. [ABSTRACT FROM AUTHOR]

The security of the Internet of Things (IoT) is a very important aspect of everyday life for people and industries, as well as hospitals, military, households and cities. Unfortunately, this topic is still too little researched and developed, which results in exposing users of Internet of Things to possible threats. One of the areas which should be addressed is the creation of a database of information about vulnerabilities and exploits in the Internet of Things; therefore, the goal of our activities under the VARIoT (Vulnerability and Attack Repository for IoT) project is to develop such a database and make it publicly available. The article presents the results of our research aimed at building this database, i.e., how the information about vulnerabilities is obtained, standardized, aggregated and correlated as well as the way of enhancing and selecting IoT related data. We have obtained and proved that existing databases provide various scopes of information and because of that a single and most comprehensive source of information does not exist. In addition, various sources present information about a vulnerability at different times-some of them are faster than others, and the differences in publication dates are significant. The results of our research show that aggregation of information from various sources can be very beneficial and has potential to enhance actionable value of information. We have also shown that introducing more sophisticated concepts, such as trust management and metainformation extraction based on artificial intelligence, could ensure a higher level of completeness of information as well as evaluate the usefulness and reliability of data.

Background: The wide adoption of social media in daily life renders it a rich and effective resource for conducting near real-time assessments of consumers' perceptions of health services. However, its use in these assessments can be challenging because of the vast amount of data and the diversity of content in social media chatter., Objective: This study aims to develop and evaluate an automatic system involving natural language processing and machine learning to automatically characterize user-posted Twitter data about health services using Medicaid, the single largest source of health coverage in the United States, as an example., Methods: We collected data from Twitter in two ways: via the public streaming application programming interface using Medicaid-related keywords (Corpus 1) and by using the website's search option for tweets mentioning agency-specific handles (Corpus 2). We manually labeled a sample of tweets in 5 predetermined categories or other and artificially increased the number of training posts from specific low-frequency categories. Using the manually labeled data, we trained and evaluated several supervised learning algorithms, including support vector machine, random forest (RF), naïve Bayes, shallow neural network (NN), k-nearest neighbor, bidirectional long short-term memory, and bidirectional encoder representations from transformers (BERT). We then applied the best-performing classifier to the collected tweets for postclassification analyses to assess the utility of our methods., Results: We manually annotated 11,379 tweets (Corpus 1: 9179; Corpus 2: 2200) and used 7930 (69.7%) for training, 1449 (12.7%) for validation, and 2000 (17.6%) for testing. A classifier based on BERT obtained the highest accuracies (81.7%, Corpus 1; 80.7%, Corpus 2) and F 1 scores on consumer feedback (0.58, Corpus 1; 0.90, Corpus 2), outperforming the second best classifiers in terms of accuracy (74.6%, RF on Corpus 1; 69.4%, RF on Corpus 2) and F 1 score on consumer feedback (0.44, NN on Corpus 1; 0.82, RF on Corpus 2). Postclassification analyses revealed differing intercorpora distributions of tweet categories, with political (400778/628411, 63.78%) and consumer feedback (15073/27337, 55.14%) tweets being the most frequent for Corpus 1 and Corpus 2, respectively., Conclusions: The broad and variable content of Medicaid-related tweets necessitates automatic categorization to identify topic-relevant posts. Our proposed system presents a feasible solution for automatic categorization and can be deployed and generalized for health service programs other than Medicaid. Annotated data and methods are available for future studies., (©Yuan-Chi Yang, Mohammed Ali Al-Garadi, Whitney Bremer, Jane M Zhu, David Grande, Abeed Sarker. Originally published in the Journal of Medical Internet Research (https://www.jmir.org), 03.05.2021.)