Past Attempts Research Articles

An explainable machine learning approach for student dropout prediction

School dropout is a relevant socio-economic problem across the globe. Predictive models have been developed to determine the likelihood of students dropping out of their studies precociously to overcome such a problem. Academic systems, which gather data from many students, are potential sources for datasets that feed dropout prediction algorithms, thus leading to general improvements in education quality. Despite successful past attempts to predict dropout, several works depict small datasets with features that are hard to reproduce. Furthermore, predicting whether a student will drop out is not enough to diagnose and prevent the problem as it is also necessary to provide potential justifications for the dropout. This paper proposes an approach for creating and enriching a dataset for dropout prediction, which has been applied for dropout prediction using data from 19 schools in Brazil. With this dataset and using classifiers and model explaining techniques, our experiments achieved Area Under the Precision–Recall Curve (AUC-PR) scores of up to 89.5%, Precision up to 95%, Recall up to 93%, and Kolmogorov–Smirnov (KS) rates up to 97% when predicting dropout at different year moments. This study also shows differences when predicting dropouts in different educational stages, such as preschool and secondary education, with the former being more complex than the latter. In addition to the high recognition rates, our proposal identifies potential reasons for student dropout, which are relevant for educational institutions to take preemptive actions.

Influence of Knowledge of Smoking as Cancer-Causing on Behavior among Smokers

Background: Smoking-related behavior including passive smoking poses a major burden on the health status of populations and national economies. This is particularly true for countries that have a high prevalence of smoking like Jordan, which also has an increasing trend of waterpipe smoking: Knowledge of passive smoking and the association of knowledge of smoking as cancer-causing with quitting in addition to characteristics of smokers with denial or ignorance were assessed.
 Methods: This study is part of a nationwide, cross-sectional survey conducted in Jordan. Respondents (3196) were asked about their knowledge of passive smoking in addition to the association of active, waterpipe, and passive smoking with cancer. Smokers were also asked about the possibility of them developing cancer. Knowledge and perception were correlated to past attempts and future intentions of quitting smoking. Smokers with a denial of developing cancer or lack of knowledge of the association of smoking with cancer were further characterized.
 Results: Approximately half of the respondents, in general, knew of the term “passive smoking.” Knowledge of the association of smoking-related behavior with cancer was very high (~95%). However, almost 75% of smokers thought that they would develop cancer with a strong association between this perception with previous attempts or future intentions to quit smoking. Whereas smokers in their 20s and 50s, those with decreasing income, and female smokers were more likely to be in denial that they may develop cancer, male smokers and smokers in their 40s, or with intermediate education or increasing income appeared to be ignorant of the association of smoking with developing cancer.
 Conclusions: These results may aid in the development of more effective and targeted anti-smoking campaigns.

Time-Deniable Signatures

In this work we propose time-deniable signatures (TDS), a new primitive that facilitates deniable authentication in protocols such as DKIM-signed email. As with traditional signatures, TDS provide strong authenticity for message content, at least {\em for a sender-chosen period of time}. Once this time period has elapsed, however, time-deniable signatures can be forged by any party who obtains a signature. This forgery property ensures that signatures serve a useful authentication purpose for a bounded time period, while also allowing signers to plausibly disavow the creation of older signed content. Most critically, and unlike many past proposals for deniable authentication, TDS do not require interaction with the receiver or the deployment of any persistent cryptographic infrastructure or services beyond the signing process ( e.g., APIs to publish secrets or author timestamp certificates.) We first investigate the security definitions for time-deniability, demonstrating that past definition attempts are insufficient (and indeed, allow for broken signature schemes.) We then propose an efficient construction of TDS based on well-studied assumptions.

The association of pain with suicidal ideation and suicide attempts with depressive symptoms among adults aged ≥50years from low- and middle-income countries.

We aimed to examine the relationship of pain with suicidal ideation and suicide attempts with depressive symptoms among adults aged ≥50years from six low- and middle-income countries (LMICs) (China, Ghana, India, Mexico, Russia, and South Africa). Cross-sectional, community-based, nationally representative data from the WHO Study on global AGEing and adult health were analyzed. Self-reported information on past 12-month suicidal ideation and suicide attempts among people with depressive symptoms was collected. Pain was assessed with the question "Overall in the last 30days, how much of bodily aches or pain did you have?" With answer options: "none", "mild", "moderate", "severe/extreme". Multivariable logistic regression was done to assess associations. Data on 34,129 adults aged ≥50years (mean [SD] age 62.4 [16.0] years; males 47.9%) were analyzed. Compared to no pain, mild, moderate, and severe/extreme pain were associated with 2.83 (95% CI=1.51-5.28), 4.01 (95% CI=2.38-6.76), and 12.26 (95% CI=6.44-23.36) times higher odds for suicidal ideation. For suicide attempt, only severe/extreme pain was associated with significantly increased odds (OR=4.68; 95% CI=1.67-13.08). In this large sample of older adults from multiple LMICs, pain was strongly associated with suicidal thoughts and suicide attempts with depressive symptoms. Future studies should assess whether addressing pain among older people in LMICs may lead to reduction in suicidal thoughts and behaviors.

Balancing globalisation process through Development Agenda: sustainable global justice

The prevailing assumption was that the phenomenon of globalisation would yield equitable benefits for all nations. Nevertheless, it has become evident that the benefits of globalisation norms are predominantly enjoyed by affluent nations. Although the concept of globalisation was initially intended to enhance the agency of developing nations, it has instead resulted in the erosion of their autonomy to independently determine their course of action. The democratic process has been impacted by the loss of decision-making power experienced by emerging countries as a consequence of globalisation. The rules governing globalisation were inequitably formulated to provide preferential treatment to developed nations. Contrary to prevailing popular opinion, the impact of globalisation on both developing and affluent nations has exhibited disparities. The questioning of norms governing globalisation has emerged as a consequence of the phenomenon of globalisation. This process further exacerbated the existing division between the global north and global south, leading to a heightened level of separation between these two regions. International financial institutions based in Washington disregarded the objections raised by developing nations and formulated a strategy aimed at mitigating the disparity in wealth. The ideology of privatisation and the free market is commonly referred to as the Washington consensus. The concept of state non-intervention proved to be disappointing as it primarily served the interests of the international organisations involved, thereby failing to benefit other stakeholders. Hence, it is imperative to formulate a development strategy that encompasses principles of equity, justice, and long-term sustainability. The existing understanding of sustainable development and the approach to attaining justice through sustainability necessitate a reevaluation in order to accomplish similar objectives. The authors of this research endeavor to investigate the extent to which the current development agenda is sufficiently comprehensive to address the shortcomings of past attempts to reconcile globalisation with development. This study aims to examine the response of the global community to the phenomenon of globalisation through the implementation of the Development Agenda. This paper examines the influence of globalisation on the policy-making processes of developing nations. Finally, the authors proceed to analyse the concepts of justice, growth, and sustainability.

The effects of doctor strikes on patient outcomes: Evidence from the English NHS

Doctor strikes cause major disruption for hospitals and patients. Past attempts to estimate impacts on patients suffer from selection issues due to changing patient composition during strikes. We address these issues by exploiting differential hospital exposure to a 2016 ‘junior’ doctor strike in England to estimate the impact of doctor strikes on patient outcomes. Using the pre-strike junior-senior doctor ratio to measure exposure, we show increased strike exposure led to larger reductions in elective volumes, but did not affect volumes, average mortality or readmission rates for emergency patients. However, greater exposure to the strike did lead to higher readmission rates for black emergency patients. This suggests that while hospitals managed to mitigate many of the negative effects of the strikes, disruptions from the strikes still had negative consequences for some minority groups.

The Need for Presurgical Evaluation for Placenta Accreta Spectrum.

Staging or grading of placenta accreta spectrum has historically relied on histopathologic evaluation of placental and uterine specimens. This approach has limited utility, since it is retrospective in nature and does not allow for presurgical planning. Here, we argue for a paradigm shift to use of clinical and imaging characteristics to define the presurgical stage. We summarize past attempts at staging, and define a new data-driven approach to determining the stage prior to delivery. Use of this model may help hospitals direct patients to the most appropriate level of care for workup and management of placenta accreta spectrum. KEY POINTS: · Staging systems that rely on histopathologic grade (accreta, increta, percreta) are unhelpful in antenatal planning for placenta accreta spectrum.. · Past attempts at pre-delivery (pre-surgical) staging have failed to account for key factors that contribute to risk and morbidity.. · We developed a data-driven model that could be easily incorporated as a decision aid into clinical practice to help clinicians decide an individual patient's risk for placenta accreta spectrum..

Identifying Neurobiological Markers in Obsessive–Compulsive Disorder: A Study Protocol for a Cross-Sectional Study in Subgroups of Differing Phenotype

Obsessive–compulsive disorder (OCD) represents a frequent and highly disabling mental disorder. Past attempts to characterize different disease subgroups focused on the time of onset (late vs. early onset), presence of insight (poor insight), and post-infectious forms (pediatric acute-onset neuropsychiatric syndrome, PANS). Each subgroup may be associated with a differing impact on cognition, functioning, sleep quality, and treatment response profile. Certain lines of evidence suggest brain-derived neurotrophic factor (BDNF) levels may differ between individuals living with OCD as compared with controls, but there is a lack of evidence on the variation of BDNF levels in OCD subgroups. Lastly, the potential of assessing inflammatory states, electroencephalogram, and polysomnography to characterize these subtypes has been hardly explored. Estimates of drug-resistance rates indicate that 20% and up to 65% of affected adults and up to 35% of the pediatric population may not benefit from pharmacological treatments. At least part of the variability in treatment response could depend on the underlying biological heterogeneity. In the present project, we aim to increase the accuracy in characterizing the phenotypical and biological signature for the different OCD subtypes through clinical, cognitive, and sleep markers, along with other possible markers that may be biologically plausible.

Content and Process of Adolescent Suicide Ideation: Implications for Risk Assessment.

This study examined differences in the content and process of suicide ideation between adolescents presenting with recent suicide ideation or a suicide attempt in clinical settings. Across two combined study samples, adolescents (N = 229; 79% female; 73% Hispanic/Latine), ages 12-19, presenting with a recent suicide attempt, recent suicide ideation with a past suicide attempt history, or recent suicide ideation with no past suicide attempt history were interviewed in detail about the process and content of their suicide ideation. The group with suicide ideation and a past suicide attempt more often reported that their recent ideation lasted greater than 4h compared tothose with suicide ideation but no past suicide attempt history. The suicide attempt group more often considered ingestion as their first method of attempt, compared to the other two suicide ideation groups, and less often considered "other" methods (e.g., jumping from a height or onto train/traffic, hanging). Wish to die was lower in the ideation-only group, compared to both other groups. Separate analyses from Study 2 suggested that the majority of adolescents' suicide ideation contained imagery; however, a higher proportion of adolescents with suicide ideation and a past suicide attempt reported imagery in their ideation than those with ideation but no past attempt. Understanding what adolescents think about when they consider suicide and how they think about it may be informative about risk of a suicide attempt.

Recurrent neural network architecture for forecasting banana prices in Gujarat, India.

The forecasting of horticulture commodity prices, such as bananas, has wide-ranging impacts on farmers, traders and end-users. The considerable volatility in horticultural commodities pricing estimates has allowed farmers to exploit various local marketplaces for profitable sales of their farm produce. Despite the demonstrated efficacy of machine learning models as a suitable substitute for conventional statistical approaches, their application for price forecasting in the context of Indian horticulture remains an area of contention. Past attempts to forecast agricultural commodity prices have relied on a wide variety of statistical models, each of which comes with its own set of limitations. Although machine learning models have emerged as formidable alternatives to more conventional statistical methods, there is still reluctance to use them for the purpose of predicting prices in India. In the present investigation, we have analysed and compared the efficacy of a variety of statistical and machine learning models in order to get accurate price forecast. Autoregressive Integrated Moving Average (ARIMA), Seasonal Autoregressive Integrated Moving Average model (SARIMA), Autoregressive Conditional Heteroscedasticity model (ARCH), Generalized Autoregressive Conditional Heteroscedasticity model (GARCH), Artificial Neural Network (ANN) and Recurrent Neural Network (RNN) were fitted to generate reliable predictions of prices of banana in Gujarat, India from January 2009 to December 2019. Empirical comparisons have been made between the predictive accuracy of different machine learning (ML) models and the typical stochastic model and it is observed that ML approaches, especially RNN, surpassed all other models in the majority of situations. Mean Absolute Percent Error (MAPE), Root Mean Square Error (RMSE), symmetric mean absolute percentage error (SMAPE), mean absolute scaled error (MASE) and mean directional accuracy (MDA) are used to illustrate the superiority of the models and RNN resulted least in terms of all error accuracy measures. RNN outperforms other models in this study for predicting accurate prices when compared to various statistical and machine learning techniques. The accuracy of other methodologies like ARIMA, SARIMA, ARCH GARCH, and ANN falls short of expectations.

Suicide Characteristics ofVeterans Hospitalized forSuicide Ideation or Attempt.

Veteran suicide rates continue to be unacceptably high, with the most common risk factor being a past suicide attempt (SA). However, some characteristics of suicidal ideation (SI) and behavior among Veterans hospitalized for suicide risk remain under-reported. One hundred and eighty-three Veterans hospitalized for either an SA or SI with intent were screened for enrollment in a treatment study to prevent suicide. Veterans completed a demographic form, the Columbia-Suicide Severity Rating Scale, and the McLean borderline personality disorder screening measure shortly after inpatient psychiatric admission. Chi-squared and t-tests were used to compare suicide characteristics (e.g., intensity, duration, deterrents, and controllability) between Veterans with and without a lifetime history of SA. Thematic analyses of the reported method of SI were conducted. Sixty-seven percent of participants were hospitalized for SI and 33% were hospitalized for SA. Twenty-one percent of Veterans hospitalized for SI also endorsed a recent SA in the weeks preceding hospitalization. Most participants reported at least one lifetime SA (71%). Veterans with a lifetime history of SA reported greater frequency and duration of ideation in the week before hospitalization (t[169] = -2.56, P = .01; t[168] = -2.04, P = .04) while also reporting that deterrents were less likely to prevent an SA (t[107.09] = -3.58, P = .001) compared to those with no lifetime SA. Overall, Veterans hospitalized for SI/SA demonstrated markers of chronic suicide risk, as most participants endorsed a past attempt in their lifetime. Some Veterans admitted for SI also reported a past month's attempt, suggesting that in certain cases, hospitalization does not immediately follow an acute suicidal crisis. A past SA differentiated Veterans on average frequency and duration of SI as well as the perception of deterrents preventing suicidal behavior. Therefore, a thorough evaluation of suicide methods and intensity may be informative in treatment planning for Veterans at greatest risk of suicide.

Use of medicines to alleviate negative emotional states among adolescents attending Special Education Centres.

Substantial research evidence indicates that adolescents commonly use a variety of pharmaceuticals. Studies in this area carried out so far in Poland have focused on youth attending mainstream schools. However, there is a lack of research covering adolescents residing in special education centres (SEC). The purpose of the study was to assess the prevalence of medicine use for negative emotional states (nervousness, difficulties in falling asleep, and depressive mood), and to identify factors associated with the use of the aforementioned medications by SEC youth. The study was conducted in 2018 on a random nationwide sample of SEC adolescents aged 12-19 (N = 1730). The logistic regression analyses included sociodemographic variables, as well as factors related to the participant's individual characteristics and to their social environment. Medicine use for negative emotional states was confirmed by about 24%-30% of respondents. The prevalence of these medicine use among SEC youth was significantly higher than among adolescents from mainstream schools. The individual risk factors associated with medicine use were gender (female), current mental and somatic health problems, past suicide attempts and psychoactive substances use while self-control turned out to be protective factor. Among factors related to the respondents' social environment, positive relationships with peers were significant protective factor. Preventive measures aimed at SEC youth should be focused on improving positive relationships with peers and developing self-control. Educational programs for youth, their parents and SEC staff to develop knowledge and skills regarding safe medicines intake would also be useful.

Association between social media use and health promotion among individuals with depression and anxiety: Insights from the 2017–2020 Health Information National Trends Survey

IntroductionIndividuals with mental disorders have elevated risk factors for cardiovascular disease, and as a result, experience earlier mortality than the general population. Yet, evidence-based strategies and interventions to promote physical and cardiometabolic health in mental health contexts remains limited. Social media has emerged as a promising and readily affordable tool to expand the reach of mental health services, including the delivery of behavioral and physical health interventions. Despite its widespread availability, the potential for social media use to support positive lifestyle behaviors in mental health contexts remains relatively unexplored. We examined the potential for using social media tools for health promotion by people with anxiety or depression. MethodsUsing cross-sectional data from the 2017–2020 edition of the Health Information National Trends Survey (HINTS), we evaluated differences between individuals with self-reported history of diagnosed depression/anxiety and the general population concerning health-related usage of social media platforms. Multivariable logistic regression was performed to evaluate the association between social media usage and positive health behaviors. ResultsOverall, individuals with mental disorders were more likely than the general population to engage in health-related social media usage. Those with anxiety or depression who engaged in health-related social media usage were significantly more likely to; report intentions to quit smoking (OR 3.13, 95% CI 1.12 – 8.70; p = 0.029) endorse past 12-month attempts at smoking cessation (OR 3.41, 95% CI 1.32 –8.82; p = 0.012) and were more likely to meet national recommendation for weekly strength training (OR 1.49, 95% CI 1.11 – 2.00; p = 0.008). ConclusionsPeople with anxiety or depression use social media for health purposes at rates higher than the general population. Our findings suggest that these tools can reach many people with mental disorders and offer a novel window of opportunity to promote physical health and positive lifestyle behavior change in this highly vulnerable population.

Review: Targeted nutrition in gestating sows: opportunities to enhance sow performance and piglet vitality.

Birth weight is a key factor for piglet survival, and therefore, there is ongoing interest in how nutrition during gestation can influence birth weight. Interestingly, sows are generally fed one single diet throughout gestation. This paper reviews past attempts to increase foetal growth to term and discusses opportunities to target nutritional manipulations at specific windows of gestation where key developmental events occur. Birth weight limits chances to survive mainly in piglets with birth weight below 1kg. These piglets represent around 16% of the population. Given the normal distribution of birth weight, the mean birth weight needs to be increased by at least 50-100g to have a meaningful impact on this proportion of the population and on perinatal survival. Based on existing variation in reported mean birth weight across a number of studies, it is argued that it is unrealistic to expect an increase in mean birth weight of more than 100g. Attempts in the past to increase birth weight have focussed on the last trimester of gestation, when foetal growth is accelerated. Increase in feed allowance or nutritional concepts that target placenta vascularisation have not been successful. It is argued that nutritional manipulations should rather focus on the middle of gestation, since in that period, placenta growth occurs and since placenta size limits the foetal size. Alternatively, nutritional manipulations can target placentation during the embryonic phase.

The persistence of blue ash in the aftermath of emerald ash borer may be due to adult oviposition preferences and reduced larval performance

Abstract We examined the health of mature blue ash, Fraxinus quadrangulata, in two forests in southwestern Ohio in relation to that of mature white ash, F. americana, and examined the potential importance of oviposition preferences and larval resistance in the persistence of blue ash. Both blue ash and white ash were largely unaffected by emerald ash borer in 2012. By 2018, nearly 90% of the blue ash trees observed in these forests had full or nearly full canopies, as opposed to less than 20% of the white ash encountered in our studies. In 2021, blue ash maintained a similar degree of canopy health as in 2018, but no standing live mature white ashes remained. Bark removals revealed no current or past larval feeding attempts in blue ash in 2018 or 2021, except for one attacked and killed tree in 2018. All white ash trees examined were attacked. In a laboratory bioassay with cut stems, emerald ash borer larvae fed less and grew significantly more slowly on blue ash than on white ash. Both reduced larval performance and reduced adult oviposition likely contribute to the persistence of blue ash in forests devastated by emerald ash borer.

Transgender and nonbinary young adults' depression and suicidality is associated with sibling and parental acceptance-rejection.

Transgender and nonbinary young adults (TNB YA) report high rates of depression and more suicidality than their cisgender counterparts. Parental rejection is a known predictor of worse mental health among TNB YA; however, less is known about TNB YA experiences of sibling acceptance-rejection. The purpose of this study was to determine how TNB YA perception of sibling and parental acceptance-rejection are related to TNB YA depression and suicidality. Cross-sectional. TNB YA (ages 18-25) who had disclosed their gender identity to an adult sibling were recruited to take part in an online study and completed measures of sibling and parent acceptance-rejection, depression, as well as lifetime and past year suicidality. Stepwise regressions were conducted to evaluate associations between acceptance-rejection and TNB YA depression and suicidality. The sample consisted of 286 TNB YA (Mage = 21.5, SD = 2.2) who were predominantly White (80.6%) and assigned female sex at birth (92.7%). Each family member's acceptance-rejection was associated with increased TNB YA depression scores when considered independently and combined. Independently, high rejection from each family member was associated with greater odds of reporting most suicidality outcomes. When all family members were considered together, only high rejection from a male parent was associated with four times greater odds of reporting lifetime suicidality. High rejection from both parents was associated with greater odds of reporting past year suicide attempt (OR: 3.26 female parent; 2.75 male parent). Rejection from family members is associated with worse depression and suicidality, and rejection from male parents may be particularly damaging. Sibling acceptance uniquely contributes to TNB YA's depression symptoms alone and in the context of parental support.

Tobacco use and cancer-related symptom burden: Analysis of the US Population Assessment of Tobacco and Health Study.

Understanding the relationship between tobacco use and symptom burden may inform tobacco treatment interventions tailored to the needs of individuals with cancer. The study included 1409 adult cancer survivors from Wave 5 of the US Food and Drug Administration Population Assessment of Tobacco and Health (PATH) Study. A multivariate analysis of variance controlling for age, sex, and race/ethnicity assessed the association of cigarette smoking and vaping on cancer-related symptom burden (fatigue, pain, emotional problems) and quality of life (QoL). Generalized linear mixed models controlling for the same factors were used to assess associations among symptom burden, QoL, and quit-smoking intentions, quit-smoking likelihood, and past 12-month smoking quit attempts. Weighted rates of current cigarette smoking and vaping were 14.21% and 2.88%, respectively. Current smoking was associated with greater fatigue (p<.0001; partial η 2 =.02), pain (p<.0001; partial η 2 =.08), emotional problems (p<.0001; partial η 2 =.02), and worse QoL (p<.0001; partial η 2 =.08). Current vaping was associated with greater fatigue (p=.001; partial η 2 =.008), pain (p=.009; partial η 2 =.005), and emotional problems (p=.04; partial η 2 =.003), but not worse QoL (p=.17). Higher cancer symptom burden was not associated with reduced interest in quitting, likelihood of quitting, or odds of past year quit attempts (p>.05 for each). Among adults with cancer, current smoking and vaping were associated with greater symptom burden. Survivors' interest in and intentions to quit smoking were not related to symptom burden. Future research should examine the role of tobacco cessation in improving symptom burden and QoL.

Comparison of the Smoking Cessation of Heated Tobacco Product Users and Conventional Cigarette Smokers in Korea.

Since the introduction of heated tobacco products (HTPs) in Korea in 2017, their annual sales have increased. Several studies have investigated the perceptions of HTPs and smoking cessation behaviors. For the first time, In 2019, questions focused on HTP use were introduced in the Korea National Health and Nutrition Examination Survey (KNHNES). This study aimed to compare smoking cessation behaviors between HTP users and conventional cigarette smokers using KNHANES data. Data of 947 current adult smokers from the 8th KNHNES (2019) were analyzed. Current smokers were divided into conventional cigarette (CC)-only, HTP-only, and dual-use groups. The general characteristics of the three groups were investigated. Differences in current intention to quit smoking and past attempts to quit smoking among the three groups were analyzed using multivariate logistic regression analysis by IBM SPSS ver. 25.0. HTP-only users demonstrated fewer future smoking cessation plans (adjusted odds ratio [AOR], 0.398; 95% confidence interval [CI], 0.195-0.813; P=0.012) and fewer smoking cessation attempts in the past year (AOR, 0.533; 95% CI, 0.298-0.954; P=0.034) than CC-only smokers. However, there was no significant difference between dual-use (CC+HTP) and CC-only smokers. While dual-use and CC-only smokers showed similar smoking cessation behaviors, HTP-only users had fewer previous attempts to quit smoking and were less likely to be currently ready to quit smoking. These findings can be explained by a decrease in the need to quit smoking due to the convenience of HTP and the perception that HTPs are less harmful than CC.

Crossed clinical features between eating disorders and types of bipolar disorder: Results from the FondaMental Advanced Centers of Expertise - Bipolar Disorder cohort

BackgroundEating disorders (EDs) are liable to alter the disease course of bipolar disorder (BD). We explored the crossed clinical features between EDs and BD, particularly as a function of BD type (BD1 vs. BD2). Methods2929 outpatients attending FondaMental Advanced Centers of Expertise were assessed for BD and lifetime EDs with a semi-structured interview, and their sociodemographic, dimensional and clinical data were collected according to a standardized procedure. For each ED type, bivariate analyses were used to investigate associations between these variables and the type of BD type followed by multinomial regressions with the variables associated with EDs and BDs after Bonferroni correction. ResultsComorbid EDs were diagnosed in 478 (16.4 %) cases, and were more prevalent in patients with BD2 than in those with BD1 (20.6 % vs. 12.4 %, p < 0.001). Regression models showed no difference according to the subtype of bipolar disorder on the characteristics of patients with anorexia nervosa (AN), bulimia nervosa (BN) or binge eating disorder (BED). After multiple adjustments, the factors differentiating BD patients with versus without ED were primarily age, gender, body mass index, more affective lability and comorbidity with anxiety disorders. BD patients with BED also scored higher regarding childhood trauma. BD patients with AN also showed higher risk of past suicide attempts than those with BED. ConclusionsIn a large sample of patients with BD, we found a high prevalence of lifetime EDs, especially for the BD2 type. EDs were associated with several severity indicators, but not with BD type-specific characteristics. This should prompt clinicians to carefully screen patients with BD for EDs, regardless of BD and ED types.

Deep Reinforcement Learning for Latency-Sensitive Communication With Adaptive Redundant Retransmissions

This paper studies packet repetition strategies over erasure channels with memory and a long feedback delay. The problem is initially formulated as a communications problem where a source wishes to transmit one message packet to a destination while minimizing both the delay and the number of transmissions. At each time instant, the sender is provided a delayed acknowledgement feedback about past attempts, and must decide whether to attempt a new transmission or not. This problem is then re-formulated as an episodic reinforcement learning problem, where an agent attempts to learn the optimal transmission policy, provided delayed feedback about past transmission attempts. The agent is helped by a channel estimator, which attempts to capture the channel memory and use that to predict probabilities of erasures in a future window. This channel estimator is also data-driven and learns the channel model without any <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a priori</i> channel knowledge. The paper presents a lower bound on the achievable trade-off between delay and number of transmissions for any channel modeled as a Markov process. Experimental results show that the combination of the proposed channel estimator and the agent can noticeably outperform naive strategies for channels with memory, and achieves results close to the lower bound.