Search

Na₁/₂Bi₁/₂TiO₃‐based materials have been earmarked for one of the first large‐volume applications of lead‐free piezoceramics in high‐power ultrasonics. Zn²⁺‐doping is demonstrated as a viable route to enhance the thermal depolarization temperature and electromechanically harden (1‐y)Na₁/₂Bi₁/₂TiO₃‐yBaTiO₃ (NBT100yBT) with a maximum achievable operating temperature of 150 °C and mechanical quality factor of 627 for 1 mole % Zn²⁺‐doped NBT6BT. Although quenching from sintering temperatures has been recently touted to enhance TF‐R, with quenching the doped compositions featuring an additional increase in TF‐R by 17 °C, it exhibits negligible effect on the electromechanical properties. The effect is rationalized considering the missing influence on conductivity and therefore, negligible changes in the defect chemistry upon quenching. High‐resolution diffraction indicates that Zn²⁺‐doped samples favor the tetragonal phase with enhanced lattice distortion, further corroborated by ²³Na Nuclear Magnetic Resonance investigations.

With the widespread of machine learning models for healthcare applications, there is increased interest in building applications for personalized medicine. Despite the plethora of proposed research for personalized medicine, very few focus on representing missingness and learning from the missingness patterns in time-series Electronic Health Records (EHR) data. The lack of focus on missingness representation in an individualized way limits the full utilization of machine learning applications towards true personalization. In this brief communication, we highlight new insights into patterns of missingness with real-world examples and implications of missingness in EHRs. The insights in this work aim to bridge the gap between theoretical assumptions and practical observations in real-world EHRs. We hope this work will open new doors for exploring directions for better representation in predictive modelling for true personalization.

The burden of diseases is rising worldwide, with unequal treatment efficacy for patient populations that are underrepresented in clinical trials. Healthcare, however, is driven by the average population effect of medical treatments and, therefore, operates in a "one-size-fits-all" approach, not necessarily what best fits each patient. These facts suggest a pressing need for methodologies to study individualized treatment effects (ITE) to drive personalized treatment. Despite the increased interest in machine-learning-driven ITE estimation models, the vast majority focus on tabular data with limited review and understanding of methodologies proposed for time-series electronic health records (EHRs). To this end, this work provides an overview of ITE works for time-series data and insights into future research. The work summarizes the latest work in the literature and reviews it in light of theoretical assumptions, types of treatment settings, and computational frameworks. Furthermore, this work discusses challenges and future research directions for ITEs in a time-series setting. We hope this work opens new directions and serves as a resource for understanding one of the exciting yet under-studied research areas.

Electronic Health Records present a valuable modality for driving personalized medicine, where treatment is tailored to fit individual-level differences. For this purpose, many data-driven machine learning and statistical models rely on the wealth of longitudinal EHRs to study patients' physiological and treatment effects. However, longitudinal EHRs tend to be sparse and highly missing, where missingness could also be informative and reflect the underlying patient's health status. Therefore, the success of data-driven models for personalized medicine highly depends on how the EHR data is represented from physiological data, treatments, and the missing values in the data. To this end, we propose a novel deep-learning model that learns the underlying patient dynamics over time across multivariate data to generate personalized realistic values conditioning on an individual's demographic characteristics and treatments. Our proposed model, IGNITE (Individualized GeNeration of Imputations in Time-series Electronic health records), utilises a conditional dual-variational autoencoder augmented with dual-stage attention to generate missing values for an individual. In IGNITE, we further propose a novel individualized missingness mask (IMM), which helps our model generate values based on the individual's observed data and missingness patterns. We further extend the use of IGNITE from imputing missingness to a personalized data synthesizer, where it generates missing EHRs that were never observed prior or even generates new patients for various applications. We validate our model on three large publicly available datasets and show that IGNITE outperforms state-of-the-art approaches in missing data reconstruction and task prediction.

In the rapidly changing healthcare landscape, the implementation of offline reinforcement learning (RL) in dynamic treatment regimes (DTRs) presents a mix of unprecedented opportunities and challenges. This position paper offers a critical examination of the current status of offline RL in the context of DTRs. We argue for a reassessment of applying RL in DTRs, citing concerns such as inconsistent and potentially inconclusive evaluation metrics, the absence of naive and supervised learning baselines, and the diverse choice of RL formulation in existing research. Through a case study with more than 17,000 evaluation experiments using a publicly available Sepsis dataset, we demonstrate that the performance of RL algorithms can significantly vary with changes in evaluation metrics and Markov Decision Process (MDP) formulations. Surprisingly, it is observed that in some instances, RL algorithms can be surpassed by random baselines subjected to policy evaluation methods and reward design. This calls for more careful policy evaluation and algorithm development in future DTR works. Additionally, we discussed potential enhancements toward more reliable development of RL-based dynamic treatment regimes and invited further discussion within the community. Code is available at https://github.com/GilesLuo/ReassessDTR., Comment: Accepted at ICML 2024. 9 pages for main content, 34 pages in total

Reinforcement learning (RL) has garnered increasing recognition for its potential to optimise dynamic treatment regimes (DTRs) in personalised medicine, particularly for drug dosage prescriptions and medication recommendations. However, a significant challenge persists: the absence of a unified framework for simulating diverse healthcare scenarios and a comprehensive analysis to benchmark the effectiveness of RL algorithms within these contexts. To address this gap, we introduce \textit{DTR-Bench}, a benchmarking platform comprising four distinct simulation environments tailored to common DTR applications, including cancer chemotherapy, radiotherapy, glucose management in diabetes, and sepsis treatment. We evaluate various state-of-the-art RL algorithms across these settings, particularly highlighting their performance amidst real-world challenges such as pharmacokinetic/pharmacodynamic (PK/PD) variability, noise, and missing data. Our experiments reveal varying degrees of performance degradation among RL algorithms in the presence of noise and patient variability, with some algorithms failing to converge. Additionally, we observe that using temporal observation representations does not consistently lead to improved performance in DTR settings. Our findings underscore the necessity of developing robust, adaptive RL algorithms capable of effectively managing these complexities to enhance patient-specific healthcare. We have open-sourced our benchmark and code at https://github.com/GilesLuo/DTR-Bench., Comment: 13 pages for main content

Advancements in wearable sensor technologies and the digitization of medical records have contributed to the unprecedented ubiquity of biomedical time series data. Data-driven models have tremendous potential to assist clinical diagnosis and improve patient care by improving long-term monitoring capabilities, facilitating early disease detection and intervention, as well as promoting personalized healthcare delivery. However, accessing extensively labeled datasets to train data-hungry deep learning models encounters many barriers, such as long-tail distribution of rare diseases, cost of annotation, privacy and security concerns, data-sharing regulations, and ethical considerations. An emerging approach to overcome the scarcity of labeled data is to augment AI methods with human-like capabilities to leverage past experiences to learn new tasks with limited examples, called few-shot learning. This survey provides a comprehensive review and comparison of few-shot learning methods for biomedical time series applications. The clinical benefits and limitations of such methods are discussed in relation to traditional data-driven approaches. This paper aims to provide insights into the current landscape of few-shot learning for biomedical time series and its implications for future research and applications., Comment: This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible

Autism spectrum disorder (ASD) is a common neurological disorder. Although the etiologies of ASD have been widely speculated, evidence also supports the pathogenic role of cutaneous inflammation in autism. The prevalence of ASD is higher in individuals with inflammatory dermatoses than in those without inflammatory diseases. Anti-inflammation therapy alleviates symptoms of ASD. Recent studies suggest a link between epidermal dysfunction and ASD. In the murine model, mice with ASD display epidermal dysfunction, accompanied by increased expression levels of proinflammatory cytokines in both the skin and the brain. Children with ASD, which develops in their early lifetime, also exhibit altered epidermal function. Interestingly, improvement in epidermal function alleviates some symptoms of ASD. This line of evidence suggests a pathogenic role of cutaneous dysfunction in ASD. Either an improvement in epidermal function or effective treatment of inflammatory dermatoses can be an alternative approach to the management of ASD. We summarize here the current evidence of the association between the skin and ASD.

Decentral booster heat pumps (BHPs) integrated in ultra-low temperature district heating (ULTDH) systems are deemed a sustainable solution for low-carbon building systems. The performance of BHPs is critical for the implementation of ULTDH systems. In this study, we experimentally investigate the effect of refrigerant charge on temperature, pressure, system coefficient of performance (COP), and heating capacity to evaluate the operability and sensitivity of ULTDH. A performance-guaranteed range of refrigerant charge is proposed to ensure the feasibility of the system. We also analyse the component-level exergy destruction and the system-level exergetic efficiency of a BHP prototype. The exergetic efficiencies range from 24.9 % to 33.4 % under operational conditions for ULTDH. The condenser and evaporator's exergy destruction is influenced by the temperature profiles on both sides, suggesting the potential to improve BHP performance by tailoring operating parameters. Additionally, the return water temperature significantly affects the system COP when the domestic hot water supply temperature is fixed. Within the allowable range of return water temperature, BHP operation with high evaporator outlet water temperature can improve the overall efficiency of the ULTDH system.

Background: Acculturation is the process of two different cultures coming into contact. It is unclear how acculturation influences Chinese immigrants’ engagement in advance care planning due to the complexity and multifaceted nature of both acculturation and advance care planning. Aims: To synthesize evidence regarding the role of Chinese immigrants’ acculturation in their engagement in advance care planning. Design: Systematic mixed-method review, registered in PROSPERO (CRD42021231822). Data sources: EMBASE, MEDLINE, Web of Science, and Google Scholar were searched for publications until January 21, 2021. Results: Twenty-one out of 1112 identified articles were included in the analysis. Of those 21 articles, 17 had a qualitative design and 13 originated from the United States. Three of four quantitative studies reported that higher acculturation levels were associated with better knowledge or higher rate of engagement in advance care planning. Analysis of qualitative studies showed that Chinese immigrants’ engagement in advance care planning was associated with their: (1) self-perceived cultural identity (native or non-native); (2) interpretation of filial piety (traditional or modern); and (3) interpretation of autonomy (individual or familial). To facilitate their engagement, Chinese immigrants prefer an implicit approach, non-family-related initiators, contextualization advance care planning in Chinese culture and using Chinese language. Conclusion: Chinese immigrants’ willingness to engage in advance care planning varied with their acculturation level. To engage them in advance care planning, we recommend adapting the introduction of advance care planning to address people’s perceptions of their cultural identity, filial piety, and autonomy, as well as their preference for certain approach, initiator, context, and language.

Background: Acculturation is the process of two different cultures coming into contact. It is unclear how acculturation influences Chinese immigrants’ engagement in advance care planning due to the complexity and multifaceted nature of both acculturation and advance care planning. Aims: To synthesize evidence regarding the role of Chinese immigrants’ acculturation in their engagement in advance care planning. Design: Systematic mixed-method review, registered in PROSPERO (CRD42021231822). Data sources: EMBASE, MEDLINE, Web of Science, and Google Scholar were searched for publications until January 21, 2021. Results: Twenty-one out of 1112 identified articles were included in the analysis. Of those 21 articles, 17 had a qualitative design and 13 originated from the United States. Three of four quantitative studies reported that higher acculturation levels were associated with better knowledge or higher rate of engagement in advance care planning. Analysis of qualitative studies showed that Chinese immigrants’ engagement in advance care planning was associated with their: (1) self-perceived cultural identity (native or non-native); (2) interpretation of filial piety (traditional or modern); and (3) interpretation of autonomy (individual or familial). To facilitate their engagement, Chinese immigrants prefer an implicit approach, non-family-related initiators, contextualization advance care planning in Chinese culture and using Chinese language. Conclusion: Chinese immigrants’ willingness to engage in advance care planning varied with their acculturation level. To engage them in advance care planning, we recommend adapting the introduction of advance care planning to address people’s perceptions of their cultural identity, filial piety, and autonomy, as well as their preference for certain approach, initiator, context, and language., Industrial Design Engineering, Methodologie en Organisatie van Design

Purpose: The family plays a major role in medical decision-making in China. Little is known about whether family caregivers understand patients’ preference for receiving life-sustaining treatments and are able to make decisions consistent with them when patients are incapable of making medical decisions. We aimed to compare preferences and attitudes concerning life-sustaining treatments of community-dwelling patients with chronic conditions and their family caregivers. Patients and Methods: We conducted a cross-sectional study among 150 dyads of community-dwelling patients with chronic conditions and their family caregivers from four communities in Zhengzhou. We measured preferences for life-sustaining treatments (cardiopulmonary resuscitation, mechanical ventilation, tube feeding, hemodialysis, chemotherapy), who should decide, the timing of making decisions, and their most important consideration. Results: The consistency of preferences for life-sustaining treatments between patients and family caregivers was poor to fair, with kappa values ranging from 0.071 for mechanical ventilation to 0.241 for chemotherapy. Family caregivers more frequently preferred each life-sustaining treatment for the patients than the patients themselves. More family caregivers than patients preferred the patient to make their own decisions about life-sustaining treatments (29% of patients and 44% of family caregivers). The most important considerations when deciding on life-sustaining treatments are family burden and the patient’s comfort and state of consciousness. Conclusion: There is a poor to fair consistency between community-dwelling older patients and their family caregivers in their preferences and attitudes towards life-sustaining treatments. A minority of patients and family caregivers preferred that patients make their own medical decisions. We recommend healthcare professionals to encourage discussions between patients and their families on future care to improve the mutu

Electronic Health Records (EHRs) contain sensitive patient information, which presents privacy concerns when sharing such data. Synthetic data generation is a promising solution to mitigate these risks, often relying on deep generative models such as Generative Adversarial Networks (GANs). However, recent studies have shown that diffusion models offer several advantages over GANs, such as generation of more realistic synthetic data and stable training in generating data modalities, including image, text, and sound. In this work, we investigate the potential of diffusion models for generating realistic mixed-type tabular EHRs, comparing TabDDPM model with existing methods on four datasets in terms of data quality, utility, privacy, and augmentation. Our experiments demonstrate that TabDDPM outperforms the state-of-the-art models across all evaluation metrics, except for privacy, which confirms the trade-off between privacy and utility., Comment: Page 2, Figure 1 is updated

Heart attack remain one of the greatest contributors to mortality in the United States and globally. Patients admitted to the intensive care unit (ICU) with diagnosed heart attack (myocardial infarction or MI) are at higher risk of death. In this study, we use two retrospective cohorts extracted from the eICU and MIMIC-IV databases, to develop a novel pseudo-dynamic machine learning framework for mortality prediction in the ICU with interpretability and clinical risk analysis. The method provides accurate prediction for ICU patients up to 24 hours before the event and provide time-resolved interpretability results. The performance of the framework relying on extreme gradient boosting was evaluated on a held-out test set from eICU, and externally validated on the MIMIC-IV cohort using the most important features identified by time-resolved Shapley values achieving AUCs of 91.0 (balanced accuracy of 82.3) for 6-hour prediction of mortality respectively. We show that our framework successfully leverages time-series physiological measurements by translating them into stacked static prediction problems to be robustly predictive through time in the ICU stay and can offer clinical insight from time-resolved interpretability

The lack of data democratization and information leakage from trained models hinder the development and acceptance of robust deep learning-based healthcare solutions. This paper argues that irreversible data encoding can provide an effective solution to achieve data democratization without violating the privacy constraints imposed on healthcare data and clinical models. An ideal encoding framework transforms the data into a new space where it is imperceptible to a manual or computational inspection. However, encoded data should preserve the semantics of the original data such that deep learning models can be trained effectively. This paper hypothesizes the characteristics of the desired encoding framework and then exploits random projections and random quantum encoding to realize this framework for dense and longitudinal or time-series data. Experimental evaluation highlights that models trained on encoded time-series data effectively uphold the information bottleneck principle and hence, exhibit lesser information leakage from trained models.

The prevalence of artificial intelligence (AI) has envisioned an era of healthcare democratisation that promises every stakeholder a new and better way of life. However, the advancement of clinical AI research is significantly hurdled by the dearth of data democratisation in healthcare. To truly democratise data for AI studies, challenges are two-fold: 1. the sensitive information in clinical data should be anonymised appropriately, and 2. AI-oriented clinical knowledge should flow freely across organisations. This paper considers a recent deep-learning advent, dataset condensation (DC), as a stone that kills two birds in democratising healthcare data. The condensed data after DC, which can be viewed as statistical metadata, abstracts original clinical records and irreversibly conceals sensitive information at individual levels; nevertheless, it still preserves adequate knowledge for learning deep neural networks (DNNs). More favourably, the compressed volumes and the accelerated model learnings of condensed data portray a more efficient clinical knowledge sharing and flowing system, as necessitated by data democratisation. We underline DC's prospects for democratising clinical data, specifically electrical healthcare records (EHRs), for AI research through experimental results and analysis across three healthcare datasets of varying data types.

External validation is often recommended to ensure the generalizability of ML models. However, it neither guarantees generalizability nor equates to a model's clinical usefulness (the ultimate goal of any clinical decision-support tool). External validation is misaligned with current healthcare ML needs. First, patient data changes across time, geography, and facilities. These changes create significant volatility in the performance of a single fixed model (especially for deep learning models, which dominate clinical ML). Second, newer ML techniques, current market forces, and updated regulatory frameworks are enabling frequent updating and monitoring of individual deployed model instances. We submit that external validation is insufficient to establish ML models' safety or utility. Proposals to fix the external validation paradigm do not go far enough. Continued reliance on it as the ultimate test is likely to lead us astray. We propose the MLOps-inspired paradigm of recurring local validation as an alternative that ensures the validity of models while protecting against performance-disruptive data variability. This paradigm relies on site-specific reliability tests before every deployment, followed by regular and recurrent checks throughout the life cycle of the deployed algorithm. Initial and recurrent reliability tests protect against performance-disruptive distribution shifts, and concept drifts that jeopardize patient safety.

Tingting Zhu,1,2 Dongling Liu,1 Agnes van der Heide,2 Ida J Korfage,2 Judith AC Rietjens2 1Department of Nursing and Health, Zhengzhou University, Zhengzhou, People’s Republic of China; 2Department of Public Health, Erasmus MC, University Medical Center Rotterdam, Rotterdam, the NetherlandsCorrespondence: Dongling Liu, Email htftah@163.comPurpose: The family plays a major role in medical decision-making in China. Little is known about whether family caregivers understand patients’ preference for receiving life-sustaining treatments and are able to make decisions consistent with them when patients are incapable of making medical decisions. We aimed to compare preferences and attitudes concerning life-sustaining treatments of community-dwelling patients with chronic conditions and their family caregivers.Patients and Methods: We conducted a cross-sectional study among 150 dyads of community-dwelling patients with chronic conditions and their family caregivers from four communities in Zhengzhou. We measured preferences for life-sustaining treatments (cardiopulmonary resuscitation, mechanical ventilation, tube feeding, hemodialysis, chemotherapy), who should decide, the timing of making decisions, and their most important consideration.Results: The consistency of preferences for life-sustaining treatments between patients and family caregivers was poor to fair, with kappa values ranging from 0.071 for mechanical ventilation to 0.241 for chemotherapy. Family caregivers more frequently preferred each life-sustaining treatment for the patients than the patients themselves. More family caregivers than patients preferred the patient to make their own decisions about life-sustaining treatments (29% of patients and 44% of family caregivers). The most important considerations when deciding on life-sustaining treatments are family burden and the patient’s comfort and state of consciousness.Conclusion: There is a poor to fair consistency between community-dwelling older pat

Survival analysis focuses on estimating time-to-event distributions which can help in dynamic risk prediction in healthcare. Extending beyond the classical Cox model, deep learning techniques have been developed which moved away from the constraining assumptions of proportional hazards. Traditional statistical models often only include static information where, in this work, we propose a novel conditional variational autoencoder-based method called DySurv, which uses a combination of static and time-series measurements from patient electronic health records to estimate the risk of death dynamically. DySurv has been tested on several time-to-event benchmarks where it outperforms existing methods, including deep learning methods, and we evaluate it on real-world intensive care unit data from MIMIC-IV and eICU. The predictive capacity of DySurv is consistent and the survival estimates remain disentangled across different datasets supporting the idea that dynamic deep learning models based on conditional variational inference in multi-task cases can be robust models for survival analysis.

Semantic communications (SC) have been expected to be a new paradigm shifting to catalyze the next generation communication, whose main concerns shift from accurate bit transmission to effective semantic information exchange in communications. However, the previous and widely-used metrics for images are not applicable to evaluate the image semantic similarity in SC. Classical metrics to measure the similarity between two images usually rely on the pixel level or the structural level, such as the PSNR and the MS-SSIM. Straightforwardly using some tailored metrics based on deep-learning methods in CV community, such as the LPIPS, is infeasible for SC. To tackle this, inspired by BERTScore in NLP community, we propose a novel metric for evaluating image semantic similarity, named Vision Transformer Score (ViTScore). We prove theoretically that ViTScore has 3 important properties, including symmetry, boundedness, and normalization, which make ViTScore convenient and intuitive for image measurement. To evaluate the performance of ViTScore, we compare ViTScore with 3 typical metrics (PSNR, MS-SSIM, and LPIPS) through 4 classes of experiments: (i) correlation with BERTScore through evaluation of image caption downstream CV task, (ii) evaluation in classical image communications, (iii) evaluation in image semantic communication systems, and (iv) evaluation in image semantic communication systems with semantic attack. Experimental results demonstrate that ViTScore is robust and efficient in evaluating the semantic similarity of images. Particularly, ViTScore outperforms the other 3 typical metrics in evaluating the image semantic changes by semantic attack, such as image inverse with Generative Adversarial Networks (GANs). This indicates that ViTScore is an effective performance metric when deployed in SC scenarios.

Recent advancements in AI applications to healthcare have shown incredible promise in surpassing human performance in diagnosis and disease prognosis. With the increasing complexity of AI models, however, concerns regarding their opacity, potential biases, and the need for interpretability. To ensure trust and reliability in AI systems, especially in clinical risk prediction models, explainability becomes crucial. Explainability is usually referred to as an AI system's ability to provide a robust interpretation of its decision-making logic or the decisions themselves to human stakeholders. In clinical risk prediction, other aspects of explainability like fairness, bias, trust, and transparency also represent important concepts beyond just interpretability. In this review, we address the relationship between these concepts as they are often used together or interchangeably. This review also discusses recent progress in developing explainable models for clinical risk prediction, highlighting the importance of quantitative and clinical evaluation and validation across multiple common modalities in clinical practice. It emphasizes the need for external validation and the combination of diverse interpretability methods to enhance trust and fairness. Adopting rigorous testing, such as using synthetic datasets with known generative factors, can further improve the reliability of explainability methods. Open access and code-sharing resources are essential for transparency and reproducibility, enabling the growth and trustworthiness of explainable research. While challenges exist, an end-to-end approach to explainability in clinical risk prediction, incorporating stakeholders from clinicians to developers, is essential for success.

Tuberculosis is a respiratory disease that is treatable with antibiotics. An increasing prevalence of resistance means that to ensure a good treatment outcome it is desirable to test the susceptibility of each infection to different antibiotics. Conventionally, this is done by culturing a clinical sample and then exposing aliquots to a panel of antibiotics, each being present at a pre-determined concentration, thereby determining if the sample isresistant or susceptible to each sample. The minimum inhibitory concentration (MIC) of a drug is the lowestconcentration that inhibits growth and is a more useful quantity but requires each sample to be tested at a range ofconcentrations for each drug. Using 96-well broth micro dilution plates with each well containing a lyophilised pre-determined amount of an antibiotic is a convenient and cost-effective way to measure the MICs of several drugs at once for a clinical sample. Although accurate, this is still an expensive and slow process that requires highly-skilled and experienced laboratory scientists. Here we show that, through the BashTheBug project hosted on the Zooniverse citizen science platform, a crowd of volunteers can reproducibly and accurately determine the MICs for 13 drugs and that simply taking the median or mode of 11-17 independent classifications is sufficient. There is therefore a potential role for crowds to support (but not supplant) the role of experts in antibiotic susceptibility testing.

Purpose: Diversity in boards has gained attention as a reflection of societal imbalances. The purpose of this paper is to investigate the impact of diversity in terms of both gender and nationality in management boards of small and medium-sized enterprises (SMEs) on firm performance from an upper echelons perspective. The authors also examine how board-specific characteristics influence the structural makeup of boards in gender and nationality diversity terms. Design/methodology/approach: The authors focus on the UK because of its individualistic society and flexible labour market and assess 309 SMEs in the manufacturing industry over 2009–2019. A 3-stage least squares (3SLS) estimator is used to analyse the data, the Shannon index to measure board diversity, return on assets as proxy for firm performance, and owner-manager presence, board member age and tenure are the board-specific characteristics of primary interest. Findings: Both gender and nationality diversity contribute to firm performance and represent distinct upper echelon characteristics that change the cognitive and psychological dynamics of boards. Firms with larger boards do not perform better, but diverse boards reduce the narrowing view of CEOs. Yet the presence of owner-managers, despite their performance-enhancing contribution, holds firms back from benefitting from diversity as a strategic choice. Originality/value: This study extends the upper echelons theory to include board diversity as an important aspect that should become more central in upper echelon thinking when understanding firm performance. The authors’ findings suggest that theoretical developments in search of understanding firm behaviour must proceed by accounting for diversity and not simply focusing on decision-making styles.

Observational studies have recently received significant attention from the machine learning community due to the increasingly available non-experimental observational data and the limitations of the experimental studies, such as considerable cost, impracticality, small and less representative sample sizes, etc. In observational studies, de-confounding is a fundamental problem of individualised treatment effects (ITE) estimation. This paper proposes disentangled representations with adversarial training to selectively balance the confounders in the binary treatment setting for the ITE estimation. The adversarial training of treatment policy selectively encourages treatment-agnostic balanced representations for the confounders and helps to estimate the ITE in the observational studies via counterfactual inference. Empirical results on synthetic and real-world datasets, with varying degrees of confounding, prove that our proposed approach improves the state-of-the-art methods in achieving lower error in the ITE estimation., Comment: accepted to AISTATS 2023

Xiao Zhao,1– 3,* Ying Wang,1– 4,* Tingting Zhu,1– 3 Huili Wu,1– 3 Diya Leng,1– 3 Zhiguo Qin,5 Yan Li,6 Daming Wu1– 3 1Department of Endodontics, the Affiliated Stomatological Hospital of Nanjing Medical University, Nanjing, People’s Republic of China; 2Jiangsu Province Key Laboratory of Oral Diseases, Nanjing, People’s Republic of China; 3Jiangsu Province Engineering Research Center of Stomatological Translational Medicine, Nanjing, People’s Republic of China; 4Nanjing Stomatological Hospital, Medical School of Nanjing University, Nanjing, People’s Republic of China; 5Nanjing Medical University, the First Clinical Medical College, Jiangsu Province Hospital, Nanjing, People’s Republic of China; 6State Key Laboratory of Bioelectronics, Jiangsu Key Laboratory for Biomaterials and Devices, School of Biological Science and Medical Engineering, Southeast University, Nanjing, People’s Republic of China*These authors contributed equally to this workCorrespondence: Daming Wu, Department of Endodontics, the Affiliated Stomatological Hospital of Nanjing Medical University, Jiangsu Province Key Laboratory of Oral Diseases, Jiangsu Province Engineering Research Center of Stomatological Translational Medicine, 1 Shanghai Road, Nanjing, 210029, People’s Republic of China, Tel +086 025-69593056, Fax +086 025-86516414, Email wdming@njmu.edu.cnBackground: Mesoporous calcium-silicate nanoparticles (MCSNs) are advanced biomaterials that have been used to control drug delivery for many years. Ultrasmall Prussian blue nanoparticles (UPBNPs) showed high peroxidase and catalase-like activities. This study evaluated the antibacterial and antibiofilm properties, mechanism and cytotoxicity of UPBNPs-MCSNs composites synthesized by both as precursors.Methods: UPBNPs-MCSNs were prepared and characterized. The antibacterial effect of UPBNPs-MCSNs was evaluated by the MTT assay and CFU counting method, and their biosafety was tested by CCK8. Then

Ultra-low temperature district heating (ULTDH) systems were proposed to operate at lower temperatures in combination with booster heat pumps working from 313 K to 333 K. This can increase the system efficiency and enable the exploitation of renewable heat sources, but only based on the proper design for system boundary conditions. The techno-economic analysis of ULTDH systems revealed that the seasonal coefficient of performance of booster heat pumps should be sufficiently high to ensure the competitiveness of ULTDH systems. We study the potential of a novel magnetocaloric heat pump integrated into a domestic hot water application based on ULTDH systems. The magnetocaloric heat pump is based on the active magnetic regenerator (AMR) system that uses the reversible magnetocaloric effect of a solid-state refrigerant with no global warming potential to build a heating/cooling cycle. The solid refrigerant can be layered with different magnetic phase-change temperatures to provide a good temperature match between the user water stream and the refrigerant flow. The heating capacity can be tuned in a wide range by combinatorial control of frequency, fluid flow rate, blow fraction and offset fraction. An experimentally validated model with an average error of ~9.2% was used to study the partial load operation. A COP improvement of 68% was obtained at a partial load of 47.4% by simultaneously tuning the frequency and fluid flow rate. Important factors for the magnetocaloric heat pump integration in the ULTDH system were identified.

Booster heat pumps (BHPs) enable the district heating (DH) system to operate at lower temperature levels, which may reduce the DH grid losses and increase the feasibility of implementing a lower temperature heat source. Compared with the existing heat pumps for producing domestic hot water, the operating conditions of BHP feature lower compression ratio and higher evaporation temperature. The performance of BHP with R134a under such conditions, especially the effect of refrigerant charge, is rarely studied. The practical knowledge of the performance sensitivity regarding the charge variation would guide the application of BHP in DH and provide charge limits and uncertainty requirements. The effect of refrigerant charge on the temperature, pressure, system Coefficient of Performance (COP), as well as the heating performance were studied experimentally. This study also demonstrates the variation of heating capacity, COP, and exergy efficiency of the BHP with respect to the temperature profile matching between water and refrigerant cycle. Experiments are mainly conducted with the inlet water temperatures at 40 ℃ on both condenser and evaporator sides. Results revealed that the discharge temperature, the suction temperature, and the subcooling are important for the system performance, which can be used as the assessment parameters of the refrigerant charge level. A performance guaranteed range for judging the appropriate charge was proposed and determined. The charge range of the BHP was recommended to be set within ±8.7 % of the optimal charge amount, where the system COP can be guaranteed higher than 95 % of the maximum COP value.

Electronic Health Records (EHRs) are a valuable asset to facilitate clinical research and point of care applications; however, many challenges such as data privacy concerns impede its optimal utilization. Deep generative models, particularly, Generative Adversarial Networks (GANs) show great promise in generating synthetic EHR data by learning underlying data distributions while achieving excellent performance and addressing these challenges. This work aims to review the major developments in various applications of GANs for EHRs and provides an overview of the proposed methodologies. For this purpose, we combine perspectives from healthcare applications and machine learning techniques in terms of source datasets and the fidelity and privacy evaluation of the generated synthetic datasets. We also compile a list of the metrics and datasets used by the reviewed works, which can be utilized as benchmarks for future research in the field. We conclude by discussing challenges in GANs for EHRs development and proposing recommended practices. We hope that this work motivates novel research development directions in the intersection of healthcare and machine learning.

We propose in this paper to exploit convolutional low density generator matrix (LDGM) codes for transmission of Bernoulli sources over binary-input output-symmetric (BIOS) channels. To this end, we present a new framework to prove the coding theorems for linear codes, which unifies the channel coding theorem, the source coding theorem and the joint source-channel coding (JSCC) theorem. In the presented framework, the systematic bits and the corresponding parity-check bits play different roles. Precisely, the noisy systematic bits are used to limit the list size of typical codewords, while the noisy parity-check bits are used to select from the list the maximum likelihood codeword. This new framework for linear codes allows that the systematic bits and the parity-check bits are transmitted in different ways and over different channels. With this framework, we prove that the Bernoulli generator matrix codes (BGMCs) are capacity-achieving over BIOS channels, entropy-achieving for Bernoulli sources, and also system-capacity-achieving for JSCC applications. A lower bound on the bit-error rate (BER) is derived for linear codes, which can be used to predict the error floors and hence serves as a simple tool to design the JSCC system. Numerical results show that the convolutional LDGM codes perform well in the waterfall region and match well with the derived error floors, which can be lowered down if required by simply increasing the encoding memory., Comment: 24 pages, 13 figures

A novel split type vacuum tube (VT) solar air thermal collection-stepped storage system (ST-VTSATC-SSS) is proposed in this study to address the problem of carbon and pollution emissions from coal-fired heating in rural areas. A VT air collector is used to decrease heat loss. Paraffin and lauric acid are used to conduct thermal stepped storage. The performance of ST-VTSATC-SSS is experimentally investigated. On the coldest day in winter in 2021(Ta = −2.1 °C), the collector efficiency is 61.7%, the collection-storage efficiency is 40.7%, and the heat storage is 13.7 MJ. During the thermal collection process, the thermal collection-storage efficiency can be improved by enhancing the air volume flow rate, it decreases with the enhancement of solar radiation, and it can reach 55.3%. During the thermal discharge process, increasing the air volume flow rate can enlarge the discharge power. The average discharge power can reach 904 W. The result reveals that ST-VTSATC-SSS can rapidly release heat. Besides, a heating application for the master bedroom is analyzed. In the heating season, this mechanism can save coal of 373.0 kg and reduce CO2 by 977.3 kg, and the heating cost is 14.6 ¥/GJ. The ST-VTSATC-SSS contributes a new, clean, and low cost heating method in rural building heating.

Members of the genus Juglans are monecious wind-pollinated trees in the family Juglandaceae with highly heterozygous genomes, which greatly complicates genome sequence assembly. The genomes of interspecific hybrids are usually comprised of haploid genomes of parental species. We exploited this attribute of interspecific hybrids to avoid heterozygosity and sequenced an interspecific hybrid Juglans microcarpa × J. regia using a novel combination of single-molecule sequencing and optical genome mapping technologies. The resulting assemblies of both genomes were remarkably complete including chromosome termini and centromere regions. Chromosome termini consisted of arrays of telomeric repeats about 8 kb long and heterochromatic subtelomeric regions about 10 kb long. The centromeres consisted of arrays of a centromere-specific Gypsy retrotransposon and most contained genes, many of them transcribed. Juglans genomes evolved by a whole-genome-duplication dating back to the Cretaceous-Paleogene boundary and consist of two subgenomes, which were fractionated by numerous short gene deletions evenly distributed along the length of the chromosomes. Fractionation was shown to be asymmetric with one subgenome exhibiting greater gene loss than the other. The asymmetry of the process is ongoing and mirrors an asymmetry in gene expression between the subgenomes. Given the importance of J. microcarpa × J. regia hybrids as potential walnut rootstocks, we catalogued disease resistance genes in the parental genomes and studied their chromosomal distribution. We also estimated the molecular clock rates for woody perennials and deployed them in estimating divergence times of Juglans genomes and those of other woody perennials.

Wild emmer (Triticum turgidum ssp. dicoccoides) is the progenitor of all modern cultivated tetraploid wheat. Its genome is large (> 10 Gb) and contains over 80% repeated sequences. The successful whole-genome-shotgun assembly of the wild emmer (accession Zavitan) genome sequence (WEW_v1.0) was an important milestone for wheat genomics. In an effort to improve this assembly, an optical map of accession Zavitan was constructed using Bionano Direct Label and Stain (DLS) technology. The map spanned 10.4 Gb. This map and another map produced earlier by us with the Bionanos Nick Label Repair and Stain (NLRS) technology were used to improve the current wild emmer assembly. The WEW_v1.0 assembly consisted of 151,912 scaffolds. Of them, 3,102 could be confidently aligned on the optical maps. Forty-seven were chimeric. They were disjoined and new scaffolds were assembled with the aid of the optical maps. The total number of scaffolds was reduced from 151,912 to 149,252 and N50 increased from 6.96 Mb to 72.63 Mb. Of the 149,252 scaffolds, 485 scaffolds, which accounted for 97% of the total genome length, were aligned and oriented on genetic maps, and new WEW_v2.0 pseudomolecules were constructed. The new pseudomolecules included 333 scaffolds (68.51 Mb) which were originally unassigned, 226 scaffolds (554.84 Mb) were placed into new locations, and 332 scaffolds (394.83 Mb) were re-oriented. The improved wild emmer genome assembly is an important resource for understanding genomic modification that occurred by domestication.

Aegilops tauschii is the donor of the D subgenome of hexaploid wheat and an important genetic resource. The reference-quality genome sequence Aet v4.0 for Ae. tauschii acc. AL8/78 was therefore an important milestone for wheat biology and breeding. Further advances in sequencing acc. AL8/78 and release of the Aet v5.0 sequence assembly are reported here. Two new optical maps were constructed and used in the revision of pseudomolecules. Gaps were closed with Pacific Biosciences long-read contigs, decreasing the gap number by 38,899. Transposable elements and protein-coding genes were reannotated. The number of annotated high-confidence genes was reduced from 39,635 in Aet v4.0 to 32,885 in Aet v5.0. A total of 2,245 biologically important genes, including those affecting plant phenology, grain quality, and tolerance of abiotic stresses in wheat was manually annotated and disease-resistance genes were annotated by a dedicated pipeline. Disease-resistance genes encoding nucleotide-binding site domains, receptor-like protein kinases, and receptor-like proteins were preferentially located in distal chromosome regions, whereas those encoding transmembrane coiled-coil proteins were dispersed more evenly along the chromosomes. Discovery, annotation, and expression analyses of microRNA (miRNA) precursors, mature miRNAs, and phasiRNAs are reported, including miRNA target genes. Other small RNAs, such as hc-siRNAs and tRFs, were characterized. These advances enhance the utility of the Ae. tauschii genome sequence for wheat genetics, biotechnology, and breeding.

Na 1/2Bi 1/2TiO 3-based materials have been earmarked for one of the first large-volume applications of lead-free piezoceramics in high-power ultrasonics. Zn 2+-doping is demonstrated as a viable route to enhance the thermal depolarization temperature and electromechanically harden (1-y)Na 1/2Bi 1/2TiO 3-yBaTiO 3 (NBT100yBT) with a maximum achievable operating temperature of 150 °C and mechanical quality factor of 627 for 1 mole % Zn 2+-doped NBT6BT. Although quenching from sintering temperatures has been recently touted to enhance T F-R, with quenching the doped compositions featuring an additional increase in T F-R by 17 °C, it exhibits negligible effect on the electromechanical properties. The effect is rationalized considering the missing influence on conductivity and therefore, negligible changes in the defect chemistry upon quenching. High-resolution diffraction indicates that Zn 2+-doped samples favor the tetragonal phase with enhanced lattice distortion, further corroborated by 23Na Nuclear Magnetic Resonance investigations., RST/Storage of Electrochemical Energy

Understanding of droplet dynamic characteristics on gas diffusion layer (GDL) of polymer electrolyte membrane (PEM) fuel cell is of great importance for cell performance improvement. This study comprehensively investigated droplet dynamic characteristics on fuel cell cathode GDL with gradient pore size distribution (GPSD). The influence of different pore sizes, GPSD and wettabilities of GDL on droplet dynamic characteristics is numerically evaluated by using the volume of fluid (VOF) method through the analysis of the interaction among the forces over droplet. Results indicate that droplet has a relatively short detachment time and a smaller detachment radius on superhydrophobic and superhydrophilic GDLs compared with on a moderately hydrophobic GDL regardless of pore size. Moreover, large pores can facilitate droplet detachment, but increase droplet detachment radius and pressure drop. Compared with uniform (U)PSD, for transverse (T)-GPSDs with a small pore distance at a relatively low airflow velocity or a large pore distance at a high airflow velocity, or for the longitudinal (L)-GPSD with a large pore in the upstream and a small pore in the downstream at a large pore distance, the droplet detachment time decreases by 34.3% and 25.2% with a reduction of pressure drop, respectively.

In this paper, the potential of expander speed regulation for optimization of ORC system adaptability to off-design working conditions has been explored through experiments. According to the obtained results, it has been found that there existed two different optimal speeds: optimal power speed (rop), optimal efficiency speed (roee/rote) which could endow the ORC system with maximum output power or highest thermoelectric/exergy efficiency. For working conditions of tw = 70, 73, 80 °C, power of expander at rop has been increased by 199.1%, 137.4%, 6.5% compared with the power output at design speed (rds). The thermoelectric efficiency at rote was 11.1%, 27.7% higher than that at rop and rds. To further investigate the influences of expander speed on the performance, energy and exergy analyses were conducted based on theoretical calculation and experiment results within the range of 400–1600 r/min. The results showed that actual exergy destruction of expander and working fluid pump was closely related to expander speed. Gaps between theoretical and actual exergy destruction rates proved that optimization of expander and working fluid pump would be more effective for the ORC system performance.

The recent availability of electronic health records (EHRs) have provided enormous opportunities to develop artificial intelligence (AI) algorithms. However, patient privacy has become a major concern that limits data sharing across hospital settings and subsequently hinders the advances in AI. Synthetic data, which benefits from the development and proliferation of generative models, has served as a promising substitute for real patient EHR data. However, the current generative models are limited as they only generate single type of clinical data for a synthetic patient, i.e., either continuous-valued or discrete-valued. To mimic the nature of clinical decision-making which encompasses various data types/sources, in this study, we propose a generative adversarial network (GAN) entitled EHR-M-GAN which simultaneously synthesizes mixed-type timeseries EHR data. EHR-M-GAN is capable of capturing the multidimensional, heterogeneous, and correlated temporal dynamics in patient trajectories. We have validated EHR-M-GAN on three publicly-available intensive care unit databases with records from a total of 141,488 unique patients, and performed privacy risk evaluation of the proposed model. EHR-M-GAN has demonstrated its superiority over state-of-the-art benchmarks for synthesizing clinical timeseries with high fidelity, while addressing the limitations regarding data types and dimensionality in the current generative models. Notably, prediction models for outcomes of intensive care performed significantly better when training data was augmented with the addition of EHR-M-GAN-generated timeseries. EHR-M-GAN may have use in developing AI algorithms in resource-limited settings, lowering the barrier for data acquisition while preserving patient privacy., Comment: Main article (22 pages, 7 figures); Appendix (15 pages, 8 figures)

The quality of crustaceans’ flesh has direct impact on consumers’ purchase choices, with water environment and dietary nutrition being effective ways to regulate flesh quality. The aim of present study was to investigate the impacts of water salinity (low, 4 and medium, 23) and dietary lipid source (fish oil and soybean oil) on nutritional values, texture, taste and odor of flesh of mud crab. While water salinity had no significant influence on nutritional values of crab flesh, crabs fed soybean oil displayed significantly lower contents of amino acids and n-3 PUFAs in muscle. However, crabs reared at low salinity showed reduced flesh hardness, chewiness and gumminess likely related to altered myofiber structure, that impacted muscle texture. Furthermore, low salinity and dietary soybean oil weakened umami taste and aroma characteristics of crab flesh associated with decreased contents of free amino acids, flavor nucleotides, inorganic ions and odor active compounds in flesh.

Employing booster heat pumps (BHP) in district heating systems can provide a way to use low temperature heat sources for district heating as they allow locally increasing the water temperature on the user side where necessary. Thereby, it can provide decreased heat loss of a district heating system. This experimental study investigated steady state thermal behavior of a BHP for cases with different forward temperature and different flow rates on the user side. The unit used R134a as working fluid. The results revealed that the Coefficient of Performance (COP) of the BHP system has an optimal value at the studied range. At a forward temperature of 45 °C the heat pump reached maximum COP of 4.95, heating capacity of 15.36 kW, and exergy efficiency of 33.4%, respectively. Lorenz COP and Lorenz efficiency is ranged from 13 to 27 and 0.16 to 0.26, respectively. Exergy destruction of each component as well as exergy efficiency of BHP used in the proposed system have also been analyzed, which showed that the compressor take the highest the exergy destruction among the components. The exergy efficiency ranged from 24.9 % to 33.4 % in the considered conditions. The share of exergy destruction of condenser and evaporator were influenced by the heat transfer condition on both sides, which reveals that the BHP can achieve potential performance improvement by better matching heat transfer profile of condenser and evaporator.

Given the abundance and ease of access of personal data today, individual privacy has become of paramount importance, particularly in the healthcare domain. In this work, we aim to utilise patient data extracted from multiple hospital data centres to train a machine learning model without sacrificing patient privacy. We develop a scheduling algorithm in conjunction with a student-teacher algorithm that is deployed in a federated manner. This allows a central model to learn from batches of data at each federal node. The teacher acts between data centres to update the main task (student) algorithm using the data that is stored in the various data centres. We show that the scheduler, trained using meta-gradients, can effectively organise training and as a result train a machine learning model on a diverse dataset without needing explicit access to the patient data. We achieve state-of-the-art performance and show how our method overcomes some of the problems faced in the federated learning such as node poisoning. We further show how the scheduler can be used as a mechanism for transfer learning, allowing different teachers to work together in training a student for state-of-the-art performance., Comment: 11 pages, 8 figures

Myocardial Infarction (MI) has the highest mortality of all cardiovascular diseases (CVDs). Detection of MI and information regarding its occurrence-time in particular, would enable timely interventions that may improve patient outcomes, thereby reducing the global rise in CVD deaths. Electrocardiogram (ECG) recordings are currently used to screen MI patients. However, manual inspection of ECGs is time-consuming and prone to subjective bias. Machine learning methods have been adopted for automated ECG diagnosis, but most approaches require extraction of ECG beats or consider leads independently of one another. We propose an end-to-end deep learning approach, DeepMI, to classify MI from normal cases as well as identifying the time-occurrence of MI (defined as acute, recent and old), using a collection of fusion strategies on 12 ECG leads at data-, feature-, and decision-level. In order to minimise computational overhead, we employ transfer learning using existing computer vision networks. Moreover, we use recurrent neural networks to encode the longitudinal information inherent in ECGs. We validated DeepMI on a dataset collected from 17,381 patients, in which over 323,000 samples were extracted per ECG lead. We were able to classify normal cases as well as acute, recent and old onset cases of MI, with AUROCs of 96.7%, 82.9%, 68.6% and 73.8%, respectively. We have demonstrated a multi-lead fusion approach to detect the presence and occurrence-time of MI. Our end-to-end framework provides flexibility for different levels of multi-lead ECG fusion and performs feature extraction via transfer learning., Comment: 10 pages

Cardiac signals, such as the electrocardiogram, convey a significant amount of information about the health status of a patient which is typically summarized by a clinician in the form of a clinical report, a cumbersome process that is prone to errors. To streamline this routine process, we propose a deep neural network capable of captioning cardiac signals; it receives a cardiac signal as input and generates a clinical report as output. We extend this further to generate multilingual reports. To that end, we create and make publicly available a multilingual clinical report dataset. In the absence of sufficient labelled data, deep neural networks can benefit from a warm-start, or pre-training, procedure in which parameters are first learned in an arbitrary task. We propose such a task in the form of discriminative multilingual pre-training where tokens from clinical reports are randomly replaced with those from other languages and the network is tasked with predicting the language of all tokens. We show that our method performs on par with state-of-the-art pre-training methods such as MLM, ELECTRA, and MARGE, while simultaneously generating diverse and plausible clinical reports. We also demonstrate that multilingual models can outperform their monolingual counterparts, informally terming this beneficial phenomenon as the blessing of multilinguality.

Jianhua Qin,1,2 Tingting Zhu,1,2 Weihua Wu,1,2 Huan Chen,3 Yi He4 1Department of Nephrology, Affiliated Hospital of Southwest Medical University, Luzhou 646000, Sichuan, People’s Republic of China; 2Sichuan Clinical Research Center for Nephropathy, Luzhou 646000, Sichuan, People’s Republic of China; 3Department of Pathogen Biology, Basic Medical College, Southwest Medical University, Luzhou 646000, Sichuan, People’s Republic of China; 4Department of Urology, Affiliated Hospital of Southwest Medical University, Luzhou 646000, Sichuan, People’s Republic of ChinaCorrespondence: Yi He Email jingyoufan46@163.comBackground: Long non-coding RNA (lncRNA) has been recognized as the new regulator and biomarker for cancers. However, in clear cell renal cell carcinoma (ccRCC), the functions of lncRNAs are not well characterized. This research aimed to probe the function of lncRNA PCED1B-AS1 in the progression of ccRCC.Materials and Methods: Quantitative real-time polymerase chain reaction (qRT-PCR) was conducted to detect the expression levels of PCED1B-AS1, microRNA-484 (miR-484), and zinc finger E-box binding homeobox 1 (ZEB1) in 40 pairs of human ccRCC tissues and corresponding adjacent kidney tissue samples. Chi-square test was employed to evaluate the association between PCED1B-AS1 expression level and clinicopathological characteristics. The effects of PCED1B-AS1, miR-484 and ZEB1 on the cell proliferation, migration and epithelial-mesenchymal transition (EMT) process of ccRCC cells were studied by CCK-8 assay, EdU cell proliferation assay, wound healing test and Western blotting. The regulatory relationships among PCED1B-AS1, miR-484, ZEB1 were examined by luciferase reporter gene assay and RNA immunoprecipitation assay.Results: PCED1B-AS1 was remarkably up-regulated in ccRCC tissues and cell lines. High expression of PCED1B-AS1 was associated with poor prognosis of the patients. Loss-of-function experiments showed that PCED1B-AS1

MicroRNAs (miRNAs) are important factors for the post-transcriptional regulation of protein-coding genes in plants and animals. They are discovered either by sequencing small RNAs or computationally. We employed a sequence-homology-based computational approach to identify conserved miRNAs and their target genes in Persian (English) walnut, Juglans regia, and its North American wild relative, J. microcarpa. A total of 119 miRNA precursors (pre-miRNAs) were detected in the J. regia genome and 121 in the J. microcarpa genome and miRNA target genes were predicted and their functional annotations were performed in both genomes. In the J. regia genome, 325 different genes were targets; 87.08% were regulated by transcript cleavage and 12.92% by translation repression. In the J. microcarpa genome, 316 different genes were targets; 88.92% were regulated by transcript cleavage and 11.08% were regulated by translation repression. Totals of 1.3% and 2.0% of all resistance gene analogues (RGA) and 2.7% and 2.6% of all transcription factors (TFs) were regulated by miRNAs in the J. regia and J. microcarpa genomes, respectively. Juglans genomes evolved by a whole genome duplication (WGD) and consist of eight pairs of fractionated homoeologous chromosomes. Within each pair, the chromosome that has more genes with greater average transcription also harbors more pre-miRNAs and more target genes than its homoeologue. While only minor differences were detected in pre-miRNAs between the J. regia and J. microcarpa genomes, about one-third of the pre-miRNA loci were not conserved between homoeologous chromosome within each genome. Pre-miRNA and their corresponding target genes showed a tendency to be collocated within a subgenome.

Accurate and reliable prediction of hospital admission location is important due to resource-constraints and space availability in a clinical setting, particularly when dealing with patients who come from the emergency department. In this work we propose a student-teacher network via reinforcement learning to deal with this specific problem. A representation of the weights of the student network is treated as the state and is fed as an input to the teacher network. The teacher network's action is to select the most appropriate batch of data to train the student network on from a training set sorted according to entropy. By validating on three datasets, not only do we show that our approach outperforms state-of-the-art methods on tabular data and performs competitively on image recognition, but also that novel curricula are learned by the teacher network. We demonstrate experimentally that the teacher network can actively learn about the student network and guide it to achieve better performance than if trained alone., Comment: 16 pages, 31 figures, In Proceedings of the 37th International Conference on Machine Learning