Structured Programming Research Articles

ASSESS THE SEPARATION ANXIETY AMONG DEAF AND MUTE STUDENTS LIVING AND STUDYING IN SPECIAL SCHOOLS

Background: Research is aimed at studying the separation anxiety among deaf and mute students living and studying in special schools due to the internalized nature of their symptoms, anxiety disorders frequently go undiagnosed in school students. Academic performance and daily activities can be severely harmed by anxiety. Additionally, untreated anxiety over time can lead to behaviors like skipping school and eventually, early school exit. Objectives of the Study:1) Assess the separation anxiety among deaf and mute students with selected demographic variables.2) To associate separation anxiety with specific demographic variables. Materials and Methods: The purpose of this descriptive study and the goal of the research was to assess the anxiety among deaf and mute students living and studying in special schools. Based on inclusion criteria and purposeful convenient sample approaches, 52 students were chosen. The information was gathered in an organized way. Results: The goal of the research was to assess the separation anxiety among deaf and mute students living and studying in special schools. The deaf and mute students represent their separation anxiety in the form of mild separation anxiety, moderate separation anxiety, and severe separation anxiety, and the result was 92.33%, 5.75%, and 1.92% respectively. Conclusion: The researchers did a descriptive research investigation on the topic as part of their undergrad curriculum to assess the separation anxiety among deaf and mute students living and studying in a special school.

The influence of onset of disease on exit from paid employment among workers in The Netherlands: A longitudinal register-based study with 9 years follow-up.

This study investigates the influence of onset of disease on exit from paid employment and whether this differs across diseases and sociodemographic groups. Register data from Statistics Netherlands on medication prescription was linked to information on employment status and demographics. Persons who were employed in 2009 and 2010 and who did not use medication for the selected disease in 2009 (N=5 889 036) were followed-up over nine years. Six diseases were identified based on medication prescription in 2010 and 2011: cardiovascular diseases, inflammatory diseases, diabetes mellitus, respiratory diseases, common mental disorders, and psychotic disorders. Four pathways out of paid employment were defined: disability benefits, unemployment, no income, and early retirement. Early exit from paid employment was defined as exiting paid employment before retirement age. Cause-specific Cox proportional hazards regression analyses were performed, with interaction terms for age, sex, and migration background. Onset of disease increased the likelihood of exit from paid employment, with the strongest associations for psychotic disorders [hazard ratio (HR) 2.88, 95% confidence interval (CI) 2.78-2.98] and common mental disorders (HR 2.00, 95% CI 1.97-2.03). Onset of disease was most strongly associated with disability benefits, followed by unemployment. The influence of common mental and psychotic disorders on disability increased until around middle-age, after which it decreased. The influence of mental health problems on exit from paid employment was stronger for persons with a non-native Dutch background and males. Onset of diseases, especially mental health disorders, is a risk for exiting paid employment before the retirement age. Effective interventions are needed to enhance an inclusive workforce and prevent involuntary loss of paid employment.

24 Persistent pulmonary hypertension and short-term neurological outcomes in infants with hypoxic ischemic encephalopathy undergoing therapeutic hypothermia

Abstract Background Neonatal hypoxic ischemic encephalopathy (HIE) after perinatal asphyxia is a serious clinical condition that may result in permanent neurologic deficits or death. There is an increased incidence of persistent pulmonary hypertension of the newborn (PPHN) in asphyxiated infants. Due to the high disease burden of HIE, studying its altered cardiopulmonary hemodynamics might identify those patients at highest risk for adverse outcome early on. Objectives To investigate the relationship between severity of PPHN and the extent of short-term brain injury in infants with moderate to severe HIE who received therapeutic hypothermia (TH). Design/Methods Retrospective cohort study of 230 TH-treated HIE infants. PPHN was defined and graded clinically and based on echocardiography findings. The primary outcome was the presence of HIE changes on brain MRI and/or death within the first month of life according to the diagnosis and severity of PPHN. Secondary outcomes were the pattern of HIE on brain MRI; abnormal electroencephalogram (EEG) background, clinical, or electrographic seizures, or need for anti-ictal treatment; clinical and laboratory differences between HIE babies with and without PPHN. A logistic regression model was performed to identify PPHN severity as risk factor for moderate-to-severe brain injury or death. Results Brain injury was not found to be different between the two groups according to presence of PPHN whether clinically or based on echocardiography. However, the combined outcome of severe brain injury or death was higher in the clinical PPHN group vs non-PPHN (32.6 vs 22.8%, p=0.014), and early exit from TH due to severe PPHN was associated with the highest mortality (71.4%). Death was found to have a relationship with the severity of PPHN based on echocardiography, with the highest incidence in severe PPHN (36.0 vs 7.7% and 10.0% in the severe, moderate, and mild PPHN groups, respectively, p=0.002). Infants with clinical PPHN were sicker with higher ventilation needs and worse laboratory values compared to those without PPHN. None of the other outcomes were different among the two groups. Conclusion In infants with HIE treated with TH, PPHN was not associated with increased risk of brain injury as evident on MRI, nor clinical or electrographic seizure burden. However, mortality was higher in the most severe PPHN categories.

The influence of onset of disease on exit from paid employment among workers in The Netherlands

Abstract Background This study investigates the influence of onset of disease on exit from paid employment, and whether this differs across diseases and sociodemographic groups. Methods Register data from Statistics Netherlands on medication prescription was linked to information on employment status and demographics. Persons who were employed in 2009 and 2010, and who did not use medication for the selected disease in 2009 (n = 5,889,036) were followed-up during 9-years. Six diseases were identified based on medication prescription in 2010 and 2011: cardiovascular diseases, inflammatory diseases, diabetes mellitus, respiratory diseases, psychological disorders, and psychotic disorders. Four pathways out of paid employment were defined: disability benefits, unemployment, no income, and early retirement. Cause-specific Cox Proportional Hazards regression analyses were performed, with interaction terms for age, sex, and migration background. Results Onset of disease increased the likelihood to exit paid employment, with strongest associations for psychotic disorders (HR 2.91, 95% CI 2.81-3.02) and psychological disorders (HR 2.01, 95%CI 1.98-2.04). Onset of disease was most strongly associated with disability benefits, followed by unemployment. The influence of psychological and psychotic disorders on disability increased until around middle-age, after which it decreased. The influence of mental health problems on exit from paid employment was stronger for persons with a non-native Dutch background and males. Conclusions Onset of diseases, especially mental health disorders, is a risk for early exit from paid employment. Effective interventions are needed to enhance an inclusive workforce and prevent involuntary loss of paid employment. Key messages • Onset of all diseases increased the likelihood of exiting paid employment, through disability benefits, followed by unemployment. • Onset of psychological and psychotic disorder had the strongest association with exiting paid employment, especially among males and workers with a migration background.

Extending the Theory of Planned Behavior – A Longitudinal Study of Entrepreneurial Intentions

Abstract Most recent research about entrepreneurial intent (EI) examines EI and its determinants as a one-time event, but whether and how EI is continuously changing during early-stage entrepreneurship has not been adequately addressed. We track the shifts in the EI of young creative sector early-stage entrepreneurs vis-à-vis their changing external circumstances and their developing perceptions of the realities of business start-up. Our findings of 11 established and 10 new determinants of EI contribute to the Theory of Planned Behavior (TPB), and guide our appreciation of EI as a dynamic, longitudinal construct, which is the link between intent and realized action. Our findings have profound implications for policy-makers and practitioners, for instance those wishing to help early-stage entrepreneurs to establish businesses when they only possess limited entrepreneurial experience. This research study, therefore, responds to calls for more longitudinal qualitative research into the determinants of EI, entrepreneurial processes, and early business exit.

FLEE: A Hierarchical Federated Learning Framework for Distributed Deep Neural Network over Cloud, Edge, and End Device

With the development of smart devices, the computing capabilities of portable end devices such as mobile phones have been greatly enhanced. Meanwhile, traditional cloud computing faces great challenges caused by privacy-leakage and time-delay problems, there is a trend to push models down to edges and end devices. However, due to the limitation of computing resource, it is difficult for end devices to complete complex computing tasks alone. Therefore, this article divides the model into two parts and deploys them on multiple end devices and edges, respectively. Meanwhile, an early exit is set to reduce computing resource overhead, forming a hierarchical distributed architecture. In order to enable the distributed model to continuously evolve by using new data generated by end devices, we comprehensively consider various data distributions on end devices and edges, proposing a hierarchical federated learning framework FLEE , which can realize dynamical updates of models without redeploying them. Through image and sentence classification experiments, we verify that it can improve model performances under all kinds of data distributions, and prove that compared with other frameworks, the models trained by FLEE consume less global computing resource in the inference stage.

318 How Are Research Associate Programs Structured? The First Updated, Cross-sectional Survey of Programs Across the US and Canada

Residencies must engage in scholarly activity and publish research, and many GME programs have developed research associate programs (RAPs) to aid with clinical research. These programs also allow students to gain clinical and research experience. No prior survey describes the characteristics of these programs and their growth across Canada and the US. This information is essential to help programs establish this method of increasing clinical research opportunities within GME.

Early retirement intentions: the impact of employment biographies, work stress and health among a baby-boomer generation

In recent years, early retirement decisions have become more frequent in the European Union despite political efforts to prevent early retirement. This is a growing problem for the social security system. The study focuses on a life course approach using employment biographies and investigates the influence of work stress and health on early retirement intentions. Data of employees who were born in either 1959 or 1965 of the German cohort study on work, age, health and work participation are analysed (n = 3338). By linking survey and register data from 1993 to 2011, a sequence analysis is conducted to identify employment biographies. To analyse the relationship between the employment biographies and intended early retirement, a longitudinal path analysis is computed and includes work stress, measured through effort-reward imbalance, and self-rated health. The statistical analyses identify three adverse employment biographies, i.e. part-time work, episodes of unemployment or marginal employment. In addition, two favourable employment biographies are determined, characterised by full-time work and few episodes of unemployment. The results of the path analysis show that employment biographies with high work-related stress have early retirement intentions. Among adverse employment biographies, indirect effects of poor health on the association between work stress and early retirement intentions are found. Unexpectedly, among full-time workers, work stress is also associated with early retirement intentions with an additional mediation through health. The findings of this study highlight the importance of the life course perspective when analysing retirement decisions. In addition to health-promoting interventions in the labour market, effects of psychosocial factors should be focussed on in order to reduce early exits from the labour market.

Fast CU Partition Decision Algorithm for VVC Intra Coding Using an MET-CNN

The newest video coding standard, the versatile video coding standard (VVC/H.266), came into effect in November 2020. Different from the previous generation standard—high-efficiency video coding (HEVC/H.265)—VVC adopts a more flexible block division structure, the quad-tree with nested multi-type tree (QTMT) structure, which improves its coding performance by 24%. However, it also causes a substantial increase in computational complexity. Therefore, this paper first proposes the concept of a stage grid map, which divides the overall division of a 32 × 32 coding unit (CU) into four stages and represents it as a structured output. Second, a multi-stage early termination convolutional neural network (MET-CNN) model is devised to predict the full partition information of a CU with a size of 32 × 32. Finally, a fast CU partition decision algorithm for VVC intra coding based on an MET-CNN is proposed. The algorithm can predict all partition information of a CU with a size of 32 × 32 and its sub-CUs in one run, completely replacing the complex rate-distortion optimization (RDO) process. It also has an early exit mechanism, thereby greatly reducing the encoding time. The experimental results illustrate that the scheme proposed in this paper reduces the encoding time by 49.24% on average, while the Bjøntegaard Delta Bit Rate (BDBR) only increases by 0.97%.

On-demand inference acceleration for directed acyclic graph neural networks over edge-cloud collaboration

As the development of deep neural network (DNN) tasks that feature a directed acyclic graph (DAG) surges in modern industries, simultaneously meeting the latency and accuracy requirements of these key DNN tasks has remained elusive. The general feasible solution is to find an optimal DNN partition that combines cloud with edge computing to implement edge–cloud collaboration for DNN inference. Nevertheless, the dynamicity of network conditions and the uncertain availability of cloud computing resources pose formidable obstacles to this collaborative inference. In this study, we formulate the optimization of DNN partition as a minimum cut problem in DAG derived from DNN and then propose an early-exit DAG-DNN inference (EDDI) framework that supports synergistically on-demand inference acceleration for DAG DNN. This framework introduces two novel components: (1) Evaluator that derives an approximately optimal solution for constructing DNN into appropriate DAG, assisting in solving the optimization problem; and (2) Optimizer that enables collaborative optimization of the early exit and DNN partition strategy at run time to improve performance while meeting user-defined latency requirements. Quantitative evaluations show that EDDI outperforms state-of-the-art schemes by 10.6%, 3.3%, and 3%, on average, in terms of model accuracy, inference latency, and throughput under diverse latency constraints, respectively. Meanwhile, latency speedup ratio increases by an average of 8% and 4% under varying network conditions and cloud server capacities, respectively.

Can workplace intervention prolong work life of older workers? A quasi-experimental study

BackgroundWe aimed to evaluate the impact of a workplace senior program intervention on early exit from labor market and on the disability retirement among older employees and work-related physical factors associated with it.MethodsA total of 259 individuals aged 55 + years participated in the study (107 in intervention and 152 were controls). A questionnaire survey was conducted among Finnish food industry employees between 2003 and 2009 and the intervention “senior program” was provided between 2004 and 2009. The type of pension for the respondents who had retired by 2019 was obtained and dichotomized as statutory vs. early labor market exit. Disability pension was investigated as a separate outcome. Information on work-related factors was obtained from the survey. Cox regression analysis was used to estimate hazard ratios (HR) with their 95% confidence intervals.ResultsFifty-one employees had early labor market exit. Of them, 70% (n = 36) were control participants. Employees in the senior program worked for longer years (mean years 7.4, 95% CI 6.4–8.1) compared to the control (6.6, 95% CI 6.3–7.5). Sixty percent lower risk of early labor market exit (HR 0.40, 95% CI 0.19–0.84) and disability pension was found among employees in the senior program compared to the control group. Good work ability had a 94% lower risk (0.06, 95% CI 0.01–0.29) of early labor market exit and 85% lower risk (0.15, 95% CI 0.03–0.73) of disability pension compared to poor work ability. Employees with musculoskeletal pain had 4 times higher risk of disability pension compared to those without musculoskeletal pain.ConclusionsA workplace senior program intervention prolonged work life and had positive effect on reducing disability pension among older industrial workers.

On-Device Saliency Prediction Based on Pseudoknowledge Distillation

Saliency prediction models aim to mimic the human visual system’s attention process, and the research has made significant progress due to recent advancements in deep convolution neural networks. However, the high memory requirements and intensive computational demands make these approaches less suitable for Internet-of-Things (IoT) devices, and there is a need for an improved computational efficiency and reduced memory footprint to facilitate distributed IoT intelligence. This article proposes a pseudoknowledge distillation (PKD) training method for creating a compact real-time saliency prediction model. The proposed method can effectively transfer knowledge from computationally expensive once-for-all (OFA-595) as a single teacher model and a combination of OFA-595 and EfficientNet-B7 as a multiteacher model to an early exit evolutionary algorithm network student model by utilizing knowledge distillation and pseudolabeling. Five saliency benchmark datasets are used to demonstrate PKD’s improved prediction performance and its reduced inference time without modifying the original student model.

Co-roosting within an enclosed space: are Chimney Swifts (Chaetura pelagica) disturbed by sharing a roost chimney with Rock Pigeons (Columba livia)

Many avian species co-roost with other avian species, but roost sharing has not previously been reported for species that roost in enclosed spaces. We report co-roosting of Chimney Swifts ( Chaetura pelagica (Linnaeus, 1758)) and Rock Pigeons ( Columba livia J.F. Gmelin, 1789) in a Sault Ste Marie, Ontario, Canada, chimney. We predicted that life history and behavioural differences between the two species would lead to disturbance in Chimney Swifts, but not Pigeons. Of two chimneys examined, Swifts exited earlier in the morning in the chimney with co-roosting than in the chimney where they roosted alone. While Pigeons appeared undisturbed by moving Swifts unless Swifts landed on them, Swifts were sensitive to Pigeon activity. In the evening, Pigeon movements caused Swifts to relocate within the chimney or to exit the chimney. Not all departed Swifts re-entered the roost following the disturbance. In the morning, Swifts appeared more tolerant of Pigeon movement, but were often disturbed by Pigeon exit or entry. Permanently exiting the roost in the evening or departing too early in the morning may increase energy expenditure costs for migrating Swifts. Installing Pigeon deterrent devices at the chimney may reduce disturbance to Chimney Swifts within the roost and potentially prevent early exit in the morning.

Leadership behaviours and health-related early exit from employment: a prospective cohort study of 55364 employees.

BackgroundAbsence of certain leadership behaviours, such as lack of feedback, recognition and involvement in employee development, has been associated with long-term sickness absence. We tested the hypothesis that absence of eight specific behaviours predicts health-related early exit from employment, and investigated differential effects in subgroups to guide future preventive initiatives.MethodsUsing Cox-proportional hazard modelling, we examined the prospective association between absence of leadership behaviours and health-related early exit from employment in a sample of 55 364 employees during 4.3 years follow-up. Leadership behaviours were measured by employee ratings in national surveys from 2012 to 2016. Exit from employment included disability pension and related measures of health-related early exit, retrieved from a national registry.ResultsWe identified 510 cases of health-related early exit from employment during follow-up. A high level of absence of leadership behaviours, was associated with an increased risk of exit from employment (hazard ratio: 1.57, 95% CI: 1.31; 1.89). Subgroup analyses showed that the association between absence of leadership behaviours and exit from employment was similar for women and men and across age groups. The association was stronger for employees with high level of education than for employees with medium/low education, and the association was not observed among employees with a prevalent depressive disorder.ConclusionsAbsence of the eight leadership behaviours is a risk factor for health-related early exit from employment in the Danish workforce. More studies are needed to confirm the results.

Risk factors for surgically treated cervical spondylosis in male construction workers: a 20-year prospective study

BACKGROUND CONTEXTDegenerative changes due to cervical spondylosis (CS) can detrimentally affect work ability and quality of life yet understanding of how physical exposure affects disease progression is limited. PURPOSETo assess the associations between occupational physical exposures and occurrence of surgically treated cervical spondylosis (ST-CS) and early exit from the labor market via disability pension. STUDY DESIGN/SETTINGProspective register study with 20 years follow-up period. PATIENT SAMPLESwedish construction workers participating in a national health surveillance project conducted between 1971 and 1993. OUTCOME MEASURESSurgically treated cervical spondylosis (ST-CS) and early labor market exit at a minimum rate of 25% time on disability pension. METHODSAssociations between occupational physical exposures (job exposure matrix) and subsequent ST-CS (National Hospital in-patient register) and early labor market exit via disability pension (Swedish Social Insurance Agency register) were assessed in a cohort of male construction workers (n=237,699). RESULTSA total of 1381 ST-CS cases were present and a 20 years incidence rate of 35.1 cases per 100,000 person years (95% confidence interval (CI) 33.2–36.9). Increased relative risk (RR) for ST-CS was found for workers exposed to non-neutral (RR 1.40, 95% CI 1.15–1.69), and awkward neck postures (1.52, 1.19–1.95), working with the hands above shoulder height (1.30, 1.06–1.60), and high upper extremity loading (1.35, 1.15–1.59). Increased risk was also present for workers who reported frequent neck (3.06, 2.18–4.30) and upper back (3.84, 2.57–5.73) pain in the 12 months prior to survey. Among workers with elevated arm exposure, higher risk was seen in those who also had more frequent neck pain. ST-CS cases took early retirement more often (41.3%) and at a younger age (53 years) than the total study cohort (14.8% and 56 years of age, respectively). CONCLUSIONSOccupational exposure to non-neutral neck postures, work with hands above shoulders and high loads born through the upper extremities increased the risk for ST-CS and early retirement due to disability. Decreasing postural and load exposure is salient for primary, secondary, and tertiary prevention of CS. Neck pain was shown to be a prognostic factor for ST-CS, which stresses the importance of acting early and taking preventative action to reduce workplace exposure, and the need for systematic medical check-ups within primary or occupational care to mitigate disease progression and early labour market exit due to disability.

Healthy worker survival effect at a high-altitude mine: prospective cohort observation

Very little is known about the factors of healthy worker survival effect at high-altitude mines. We conducted this cohort observation of the new hires for a high-altitude gold mine in Kyrgyzstan with the aim to ascertain predictors of survival at work. All new hires in 2009 through 2012 for a high-altitude gold mine (3600–4500 m above sea level) were followed up to January 2022. We tested the association of demographic, physiological predictors and diagnoses at the pre-employment screening with non-survival at work in Cox proportional hazards yielding hazard ratios (HR) with their 95% confidence intervals (CI). The cumulative observation time was 5190 person-years. Blood pressure at pre-employment, lung function, the diagnoses of essential hypertension, chronic obstructive pulmonary disease (COPD) or any other analyzed physiological variables were not associated with non-survival at work. However, smoking (HR 1.55; 95% CI 1.10; 2.17) increased the likelihood of non-survival at work, independent of any diagnosis or lowland residence (HR 1.95; 95% CI 1.31; 2.90). Adjusted for covariates and all diagnoses, having chronic rheumatic fever (HR 10.95; 95% CI 2.92; 33.92), hemorrhoids (HR 1.32; 95% CI 1.01; 3.75), adhesive otitis (HR 1.74; 95% CI 1.05; 2.89) or obesity (HR 1.71; 95% CI 1.01; 2.88) were associated with non-survival at work with time. This prospective observation of new hires for a high-altitude mining operation demonstrated that selected diagnoses, smoking and lowland residence elevated the risk of early exit in prospective workers.

Teaching Multiple School Subjects Role Conflict: A Theoretically Informed Conceptual Framework

Emerging research suggests that the stress and complexities of the teaching profession contribute to early exits from the field. Stressors may be increased when individuals are tasked with teaching physical education and another school subject(s) concurrently. More specifically, role conflict in teaching multiple school subjects consists of three subdomains: status conflict, schedule conflict, and energy expenditure. The purpose of this paper is to propose a theoretically informed conceptual model of this type of conflict that better informs the professional lives and careers of teachers. The framework’s three interrelated elements are dynamic and contextually bound and influence the experience of multiple subjects role conflict. These three elements include experiences of role conflict, contextual and individual factors, and an outer limit of individuals’ capacity to manage stressors. Three vignettes are used to illustrate how teachers’ experiences of conflict interact with contextual and individual factors to increase or decrease their capacity for stress.

Connections, community and context: The importance of post-boarding school pathways and re-engagement for remote Aboriginal students

For many remote Aboriginal Australian students, periods of time during their secondary education are spent living away from home at a boarding school. While financial, political and community support is burgeoning for boarding models that provide scholarships, sports programs or accommodation for Aboriginal and Torres Strait Islander students, very little academic research or evidence exists that examines the experiences of students post-boarding. This paper forms part of a broader doctoral research study, but specifically focuses on how past students, families and communities from remote South Australia view the outcomes of boarding. Using a Grounded theory design, thematic analysis of 32 semi-structured interviews with past students, families and community members led to the identification of three main themes: connections (early exits), community (re-engaging in education), and context (employment in remote communities). Findings indicated that outcomes are not linear nor easily defined. Developing a theory of change was recommended as a future approach to help families, students and remote schools to clearly define goals and measures of success for each student, recognising a range of interpretations and conceptions of ‘success’, and adapting these goals as necessary.

Towards Adversarial Robustness with Early Exit Ensembles.

Deep learning techniques are increasingly used for decision-making in health applications, however, these can easily be manipulated by adversarial examples across different clinical domains. Their security and privacy vulnerabilities raise concerns about the practical deployment of these systems. The number and variety of the adversarial attacks grow continuously, making it difficult for mitigation approaches to provide effective solutions. Current mitigation techniques often rely on expensive re-training procedures as new attacks emerge. In this paper, we propose a novel adversarial mitigation technique for biosignal classification tasks. Our approach is based on recent findings interpreting early exit neural networks as an ensemble of weight sharing sub-networks. Our experiments on state-of-the-art deep learning models show that early exit ensembles can provide robustness generalizable to various white box and universal adversarial attacks. The approach increases the accuracy of vulnerable deep learning models up to 60 percentage points, while providing adversarial mitigation comparable to adversarial training. This is achieved without previous exposure to the adversarial perturbation or the computational burden of re-training.

Template Matching Based Early Exit CNN for Energy-efficient Myocardial Infarction Detection on Low-power Wearable Devices

Myocardial Infarction (MI), also known as heart attack, is a life-threatening form of heart disease that is a leading cause of death worldwide. Its recurrent and silent nature emphasizes the need for continuous monitoring through wearable devices. The wearable device solutions should provide adequate performance while being resource-constrained in terms of power and memory. This paper proposes an MI detection methodology using a Convolutional Neural Network (CNN) that outperforms the state-of-the-art works on wearable devices for two datasets - PTB and PTB-XL, while being energy and memory-efficient. Moreover, we also propose a novel Template Matching based Early Exit (TMEX) CNN architecture that further increases the energy efficiency compared to baseline architecture while maintaining similar performance. Our baseline and TMEX architecture achieve 99.33% and 99.24% accuracy on PTB dataset, whereas on PTB-XL dataset they achieve 84.36% and 84.24% accuracy, respectively. Both architectures are suitable for wearable devices requiring only 20 KB of RAM. Evaluation of real hardware shows that our baseline architecture is 0.6x to 53x more energy-efficient than the state-of-the-art works on wearable devices. Moreover, our TMEX architecture further improves the energy efficiency by 8.12% (PTB) and 6.36% (PTB-XL) while maintaining similar performance as the baseline architecture.