Expert Performance Research Articles

AI in the ED: Assessing the efficacy of GPT models vs. physicians in medical score calculation

Background and aimsArtificial Intelligence (AI) models like GPT-3.5 and GPT-4 have shown promise across various domains but remain underexplored in healthcare. Emergency Departments (ED) rely on established scoring systems, such as NIHSS and HEART score, to guide clinical decision-making. This study aims to evaluate the proficiency of GPT-3.5 and GPT-4 against experienced ED physicians in calculating five commonly used medical scores. MethodsThis retrospective study analyzed data from 150 patients who visited the ED over one week. Both AI models and two human physicians were tasked with calculating scores for NIH Stroke Scale, Canadian Syncope Risk Score, Alvarado Score for Acute Appendicitis, Canadian CT Head Rule, and HEART Score. Cohen's Kappa statistic and AUC values were used to assess inter-rater agreement and predictive performance, respectively. ResultsThe highest level of agreement was observed between the human physicians (Kappa = 0.681), while GPT-4 also showed moderate to substantial agreement with them (Kappa values of 0.473 and 0.576). GPT-3.5 had the lowest agreement with human scorers. These results highlight the superior predictive performance of human expertise over the currently available automated systems for this specific medical outcome. Human physicians achieved a higher ROC-AUC on 3 of the 5 scores, but none of the differences were statistically significant. ConclusionsWhile AI models demonstrated some level of concordance with human expertise, they fell short in emulating the complex clinical judgments that physicians make. The study suggests that current AI models may serve as supplementary tools but are not ready to replace human expertise in high-stakes settings like the ED. Further research is needed to explore the capabilities and limitations of AI in emergency medicine.

Automated echocardiographic left ventricular dimension assessment in dogs using artificial intelligence: Development and validation.

Artificial intelligence (AI) could improve accuracy and reproducibility of echocardiographic measurements in dogs. A neural network can be trained to measure echocardiographic left ventricular (LV) linear dimensions in dogs. Training dataset: 1398 frames from 461 canine echocardiograms from a single specialist center. 50 additional echocardiograms from the same center. Training dataset: a right parasternal 4-chamber long axis frame from each study, labeled by 1 of 18 echocardiographers, marking anterior and posterior points of the septum and free wall. End-diastolic and end-systolic frames from 50 studies, annotated twice (blindly) by 13 experts, producing 26 measurements of each site from each frame. The neural network also made these measurements. We quantified its accuracy as the deviation from the expert consensus, using the individual-expert deviation from consensus as context for acceptable variation. The deviation of the AI measurement away from the expert consensus was assessed on each individual frame and compared with the root-mean-square-variation of the individual expert opinions away from that consensus. For the septum in end-diastole, individual expert opinions deviated by 0.12 cm from the consensus, while the AI deviated by 0.11 cm (P = .61). For LVD, the corresponding values were 0.20 cm for experts and 0.13 cm for AI (P = .65); for the free wall, experts 0.20 cm, AI 0.13 cm (P < .01). In end-systole, there were no differences between individual expert and AI performances. An artificial intelligence network can be trained to adequately measure linear LV dimensions, with performance indistinguishable from that of experts.

Inner sense of rhythm: percussionist brain activity during rhythmic encoding and synchronization.

The main objective of this research is to explore the core cognitive mechanisms utilized by exceptionally skilled percussionists as they navigate complex rhythms. Our specific focus is on understanding the dynamic interactions among brain regions, respectively, related to externally directed cognition (EDC), internally directed cognition (IDC), and rhythm processing, defined as the neural correlates of rhythm processing (NCRP). The research involved 26 participants each in the percussionist group (PG) and control group (CG), who underwent task-functional magnetic resonance imaging (fMRI) sessions focusing on rhythm encoding and synchronization. Comparative analyses were performed between the two groups under each of these conditions. Rhythmic encoding showed decreased activity in EDC areas, specifically in the right calcarine cortex, left middle occipital gyrus, right fusiform gyrus, and left inferior parietal lobule, along with reduced NCRP activity in the left dorsal premotor, right sensorimotor cortex, and left superior parietal lobule. During rhythmic synchronization, there was increased activity in IDC areas, particularly in the default mode network, and in NCRP areas including the left inferior frontal gyrus and bilateral putamen. Conversely, EDC areas like the right dorsolateral prefrontal gyrus, right superior temporal gyrus, right middle occipital gyrus, and bilateral inferior parietal lobule showed decreased activity, as did NCRP areas including the bilateral dorsal premotor cortex, bilateral ventral insula, bilateral inferior frontal gyrus, and left superior parietal lobule. PG's rhythm encoding is characterized by reduced cognitive effort compared to CG, as evidenced by decreased activity in brain regions associated with EDC and the NCRP. Rhythmic synchronization reveals up-regulated IDC, down-regulated EDC involvement, and dynamic interplay among regions with the NCRP, suggesting that PG engages in both automatic and spontaneous processing simultaneously. These findings provide valuable insights into expert performance and present opportunities for improving music education.

Optimizing organizational success: Matching the right position with the right person

The purpose of this study is to identify the core capabilities required of Information Technology workers working in IT firms. Personal competencies, technical competencies, job-related competencies, interpersonal competencies, knowledge level competencies, and management competencies were all employed in the study, and they are all linked to IT employees’ performance and help them thrive in their jobs. The existing study is primary in nature and involves retrieval of responses from IT employees through structured questionnaire ranged from 1–5 Likert scale. The sample size of the study is 176 respondents using purposive random sampling techniques. The study ‘s findings imply that in order to get a competitive advantage in the business industry, every individual in the company should be involved in improving productivity and efficiency at work. Similarly, the firm faces a difficult task in retaining people and persuading them to stay current in their fields of expertise or critical performance areas. As a result, competency mapping will aid the organisation in finding the expected skills of each employee in relation to the jobʼs actual skill set. Specific training initiatives to overcome the skill shortfall of IT employees should be taken based on the identified gaps.

Peak occipital alpha frequency mediates the relationship between sporting expertise and multiple object tracking performance.

Multiple object tracking (MOT) is often used as a lab-based paradigm for investigating goal-driven attention as an indicator for "real-world" attention in tasks such as sport. When exploring MOT performance in the context of sporting expertise, we typically observe that individuals with sporting expertise outperform non-sporting individuals. There are a number of general explanations for performance differences such as cognitive transfer effects; however, the potential neurophysiological mechanisms explaining the relationship between sporting expertise and performance differences in MOT are not clear. Based on the role occipital alpha (posterior oscillations usually around 8-12Hz) has been shown to have in visuospatial attention, the aim of this study was to examine whether individual differences in occipitalpeak alpha frequency (PAF) mediate the relationship between sporting expertise and performance in two object tracking tasks: a standard MOT task and a visuomotor-controlled object tracking task (multiple object avoidance [MOA]). Using electroencephalography (EEG), participants, who either played sport competitively or did not, had their posterior PAF measured at rest (eyes closed) across a 2-min window. They completed the two tasks separately from the restingEEGmeasures. Those who engaged in sport performed better in the MOT and MOA tasks and had higher PAF. Higher PAF predicted superior MOT performance. The mediation analysis revealed that sporting individuals had significantly higher PAF, and this was in turn related to superior MOT performance. It is suggested that PAF is a possible neurophysiological mediating mechanism as to why sporting individuals have superior MOT performance. There was no evidence that PAF mediated the relationship between sporting expertise and visuomotor MOA performance. Explanations and implications are discussed, and unanswered questions are proposed.

Retinal OCT Image Classification Based on CNN-RNN Unified Neural Networks

&lt;p&gt;Computer-aided diagnosis of retinopathy is a hot research topic in the field of medical image classification, where optical coherence tomography (OCT) is an important basis for the diagnosis of ophthalmic diseases. Traditional approaches to multi-label image classification learn independent classifiers for each category and employ ranking or thresholding on the classification results. These techniques, although working well, fail to explicitly exploit the label dependencies in an image. In this paper, two publicly available retinal OCT image datasets are integrated and screened. Then, an end-to-end deep learning algorithmic framework based on CNN-RNN Unified Neural Networks was proposed to automatically and reliably classify six categories of retinal OCT images. Numerical results suggest that the proposed algorithm works well in terms of accuracy, precision, sensitivity and specificity, approaching or even partially surpassing the performance of clinical experts. It is valuable in promoting computer-aided diagnosis towards practical clinical applications and improving the efficiency of clinical diagnosis of retinal diseases.&lt;/p&gt; &lt;p&gt;&amp;nbsp;&lt;/p&gt;

Observing an expert's action swapped with an observer's face increases corticospinal excitability during combined action observation and motor imagery.

This study aimed to examine whether observing an expert's action swapped with an observer's face increases corticospinal excitability during combined action observation and motor imagery (AOMI). Twelve young males performed motor imagery of motor tasks with different difficulties while observing the actions of an expert performer and an expert performer with a swapped face. Motor tasks included bilateral wrist dorsiflexion (EASY) and unilateral two-ball rotating motions (DIFF). During the AOMI of EASY and DIFF, single-pulse transcranial magnetic stimulation was delivered to the left primary motor cortex, and motor-evoked potentials (MEPs) were obtained from the extensor carpi ulnaris and first dorsal interosseous muscles of the right upper limb, respectively. Visual analogue scale (VAS) assessed the subjective similarity of the expert performer with the swapped face in the EASY and DIFF to the participants themselves. The MEP amplitude in DIFF was larger in the observation of the expert performer with the swapped face than that of the expert performer (P = 0.012); however, the corresponding difference was not observed in EASY (P = 1.000). The relative change in the MEP amplitude from observing the action of the expert performer to that of the expert performer with the swapped face was positively correlated with VAS only in DIFF (r = 0.644, P = 0.024). These results indicate that observing the action of an expert performer with the observer's face enhances corticospinal excitability during AOMI, depending on the task difficulty and subjective similarity between the expert performer being observed and the observer.

Altered spontaneous regional brain activity in ventromedial prefrontal cortex and visual area of expert table tennis athletes.

Sports training may lead to functional changes in the brain, and different types of sports, including table tennis, have different influences on these changes. However, the effects of long-term table tennis practice on brain function in expert athletes are largely undefined. Here, we investigated spontaneous regional brain activity characteristics of expert table tennis athletes by using resting-state functional magnetic resonance imaging to compare differences between 25 athletes and 33 age- and sex-matched non-athletes. We analyzed four metrics-amplitude of low-frequency fluctuation (ALFF), fractional ALFF, regional homogeneity, and degree centrality-because together they identify functional changes in the brain with greater sensitivity than a single indicator and may more comprehensively describe regional functional changes. Additional statistical analysis was conducted to assess whether any correlation existed between brain activity and years of table tennis training for athletes. Results show that compared with non-athletes, table tennis athletes showed altered spontaneous regional brain activity in the ventromedial prefrontal cortex and the calcarine sulcus, a visual area. Furthermore, the functional changes in the calcarine sulcus showed a significant correlation with the number of years of expert sports training. Despite the relatively small sample size, these results indicated that the regional brain function of table tennis athletes was associated with sports training-related changes, providing insights for understanding the neural mechanisms underlying the expert performance of athletes.

“I read the rules and know what is expected of me”: The performance of competence and expertise in ‘newbie’ offenders’ membership requests to dark web child abuse communities

Community-building among groups of child abusers on the ‘dark web’ facilitates the large-scale distribution of indecent imagery and supports individuals in becoming more skilled, more dangerous offenders. Undercover police are tasked with posing as offenders to gather intelligence; however, we know little about the nature of these groups, and especially how one might approach them linguistically as an ‘authentically’ interested outsider. This study analyses rhetorical moves (Swales, 1990) in forum posts from child abuse-related dark web fora by self-identifying ‘newbies’ hoping to join established abuse communities. It identifies 12 distinct moves used in the pursuit to join online abuse communities and finds that expressions of competence and expertise are central to newbies’ attempts to gain community membership. ‘De-lurking’ is identified as a useful strategy in the performance of competence in online forums. These findings can support online undercover policing tasks as well as offender prioritisation.

EEG-based neurophysiological indices for expert psychomotor performance – a review

A primary objective of current human neuropsychological performance research is to define the physiological correlates of adaptive knowledge utilization, in order to support the enhanced execution of both simple and complex tasks. Within the present article, electroencephalography-based neurophysiological indices characterizing expert psychomotor performance, will be explored. As a means of characterizing fundamental processes underlying efficient psychometric performance, the neural efficiency model will be evaluated in terms of alpha-wave-based selective cortical processes. Cognitive and motor domains will initially be explored independently, which will act to encapsulate the task-related neuronal adaptive requirements for enhanced psychomotor performance associating with the neural efficiency model. Moderating variables impacting the practical application of such neuropsychological model, will also be investigated. As a result, the aim of this review is to provide insight into detectable task-related modulation involved in developed neurocognitive strategies which support heightened psychomotor performance, for the implementation within practical settings requiring a high degree of expert performance (such as sports or military operational settings).

Detection of Bowel and Mesenteric Injuries Using Deep Learning Computer Vision Models

Background/Objective:While only seen in 1-5% of patients who undergo a CT (computed tomography) scan, blunt bowel and mesenteric injuries (BMI) are associated with significantly increased morbidity and mortality. A significant cause of the increased morbidity of BMI is due to the difficulty of diagnosis from clinical and imaging information which leads to delay in diagnosis. Accurate and timely diagnosis is vital to reduce the morbidity of BMI.&#x0D; Methods:For this project, our primary objective is to create a binary prediction model that determines if a patient has BMI based on their abdominal CT scans. Due to the importance of the early and definitive diagnosis of BMI in trauma patients, an extension of this project will seek to introduce explainability into the model to highlight which features on the CT scan caused the model to make its prediction. The patients with BMI were sourced from a trauma registry that recorded trauma cases from IU Health with relevant diagnosis codes. The images from our search will be reduced to the relevant slices for diagnosis of BMI and then used to train an ML model to makea yes/no prediction from the image. Once the model is trained, testing data will be evaluated on the model and the gradient vectors from the model during inference will be used to create a heatmap with GRAD-CAM that illustrates what portions of the image were relevant for the decision made by the algorithm.&#x0D; Future Directions:Using the collected abdominal CTs, we can train our machine learning pipeline to detect BMI. Based on the performance of the model, we will determine if we need to collect more data. Then, we can evaluate the explainability of the model using GRAD-CAM and compareperformance of the ML model to the performance of expert and trainee radiologists.

Approaching expert-level accuracy for differentiating ACL tear types on MRI with deep learning

Treatment for anterior cruciate ligament (ACL) tears depends on the condition of the ligament. We aimed to identify different tear statuses from preoperative MRI using deep learning-based radiomics with sex and age. We reviewed 862 patients with preoperative MRI scans reflecting ACL status from Hunan Provincial People’s Hospital. Based on sagittal proton density-weighted images, a fully automated approach was developed that consisted of a deep learning model for segmenting ACL tissue (ACL-DNet) and a deep learning-based recognizer for ligament status classification (ACL-SNet). The efficacy of the proposed approach was evaluated by using the sensitivity, specificity and area under the receiver operating characteristic curve (AUC) and compared with that of a group of three orthopedists in the holdout test set. The ACL-DNet model yielded a Dice coefficient of 98% ± 6% on the MRI datasets. Our proposed classification model yielded a sensitivity of 97% and a specificity of 97%. In comparison, the sensitivity of alternative models ranged from 84 to 90%, while the specificity was between 86 and 92%. The AUC of the ACL-SNet model was 99%, demonstrating high overall diagnostic accuracy. The diagnostic performance of the clinical experts as reflected in the AUC was 96%, 92% and 88%, respectively. The fully automated model shows potential as a highly reliable and reproducible tool that allows orthopedists to noninvasively identify the ACL status and may aid in optimizing different techniques, such as ACL remnant preservation, for ACL reconstruction.

End-to-end volumetric segmentation of white matter hyperintensities using deep learning

Background and objectivesReliable detection of white matter hyperintensities (WMH) is crucial for studying the impact of diffuse white-matter pathology on brain health and monitoring changes in WMH load over time. However, manual annotation of 3D high-dimensional neuroimages is laborious and can be prone to biases and errors in the annotation procedure. In this study, we evaluate the performance of deep learning (DL) segmentation tools and propose a novel volumetric segmentation model incorporating self-attention via a transformer-based architecture. Ultimately, we aim to evaluate diverse factors that influence WMH segmentation, aiming for a comprehensive analysis of the state-of-the-art algorithms in a broader context. MethodsWe trained state-of-the-art DL algorithms, and incorporated advanced attention mechanisms, using structural fluid-attenuated inversion recovery (FLAIR) image acquisitions. The anatomical MRI data utilized for model training was obtained from healthy individuals aged 62–70 years in the Live active Successful Aging (LISA) project. Given the potential sparsity of lesion volume among healthy aging individuals, we explored the impact of incorporating a weighted loss function and ensemble models. To assess the generalizability of the studied DL models, we applied the trained algorithm to an independent subset of data sourced from the MICCAI WMH challenge (MWSC). Notably, this subset had vastly different acquisition parameters compared to the LISA dataset used for training. ResultsConsistently, DL approaches exhibited commendable segmentation performance, achieving the level of inter-rater agreement comparable to expert performance, ensuring superior quality segmentation outcomes. On the out of sample dataset, the ensemble models exhibited the most outstanding performance. ConclusionsDL methods generally surpassed conventional approaches in our study. While all DL methods performed comparably, incorporating attention mechanisms could prove advantageous in future applications with a wider availability of training data. As expected, our experiments indicate that the use of ensemble-based models enables the superior generalization in out-of-distribution settings. We believe that introducing DL methods in the WHM annotation workflow in heathy aging cohorts is promising, not only for reducing the annotation time required, but also for eventually improving accuracy and robustness via incorporating the automatic segmentations in the evaluation procedure.

Evaluation of the Impact of Hybrid Working on Auditors' Self-Efficacy

This study aims to evaluate the concept of hybrid working in Public Accounting Firms (KAP) and its impact on self-efficacy after the COVID-19 pandemic who previously worked remotely. This research was conducted using a qualitative research method. The data source of this research was carried out by distributing questionnaires with 53 respondents from associate to manager level who work at Public Accounting Firm and interviews with 5 resource persons from senior associate to manager level from Big Four and Non-Big Four Firms. The results showed that in implementing the hybrid concept, Firms has disseminated the Firms' policy, provided support, provided facilities for collaboration and focus, as well as provided flexibility in determining work schedules. Self-efficacy arising from hybrid working shows experience, expertise, confidence, motivation, and auditor performance help auditors adapt to hybrid work. For this reason, this study recommends that the hybrid working model can be continued as a new working trend of auditors in KAP. This research is still limited to auditors below manager level and does not represent all KAP in Indonesia. This research provides an understanding of the concept of hybrid working which is a new work style choice in the auditor environment by looking at the auditor's personal self-efficacy.

International Sign conference interpreters as a Community of Practice

Conference signed language interpreters traditionally work between their national sign language and the national spoken or written language. During their careers, interpreters may add other spoken or signed languages to their repertoire. Some also acquire International Sign (IS) as a working language and interpret between IS and other signed and spoken languages. Being highly context dependent, IS has limited conventions and there is no established educational path toward learning to interpret IS. Generally, the acquisition of IS by any signer happens through interaction with signers of other signed languages. In this article, we explore the concept of IS conference interpreters as a Community of Practice (CoP), where novices acquire and experienced IS interpreters further their IS interpreting competences through situated learning. Such learning in practice may ultimately lead to the development of interpreting expertise and expert performance in IS conference interpreting. We present new data from a 2019 global survey of IS conference interpreters and follow-up interviews with eleven selected survey respondents. The results of our study suggest that there is indeed a CoP of IS conference interpreters and that it is essential for individual interpreters to participate in that community to develop the required competences and professional practices.

Image segmentation of impacted mesiodens using deep learning.

This study aimed to evaluate the performance of deep learning algorithms for the classification and segmentation of impacted mesiodens in pediatric panoramic radiographs. A total of 850 panoramic radiographs of pediatric patients (aged 3-9 years) was included in this study. The U-Net semantic segmentation algorithm was applied for the detection and segmentation of mesiodens in the upper anterior region. For enhancement of the algorithm, pre-trained ResNet models were applied to the encoding path. The segmentation performance of the algorithm was tested using the Jaccard index and Dice coefficient. The diagnostic accuracy, precision, recall, F1-score and time to diagnosis of the algorithms were compared with those of human expert groups using the test dataset. Cohen's kappa statistics were compared between the model and human groups. The segmentation model exhibited a high Jaccard index and Dice coefficient (>90%). In mesiodens diagnosis, the trained model achieved 91-92% accuracy and a 94-95% F1-score, which were comparable with human expert group results (96%). The diagnostic duration of the deep learning model was 7.5 seconds, which was significantly faster in mesiodens detection compared to human groups. The agreement between the deep learning model and human experts is moderate (Cohen's kappa = 0.767). The proposed deep learning algorithm showed good segmentation performance and approached the performance of human experts in the diagnosis of mesiodens, with a significantly faster diagnosis time.

Expertise-dependent visuocognitive performance of chess players in mating tasks: evidence from eye movements during task processing.

Visuocognitive performance is closely related to expertise in chess and has been scrutinized by several investigations in the last decades. The results indicate that experts' decision-making benefits from the chunking process, perception and visual strategies. Despite numerous studies which link these concepts, most of these investigations have employed common research designs that do not use real chess play, but create artificial laboratory conditions via screen-based chess stimuli and obtrusive stationary eye tracking with or without capturing of decision-making or virtual reality settings. The present study assessed the visuocognitive performance of chess novices, intermediates and experts in a real chess setting. Instead of check detection, find-the-best-move tasks or to distinguish between regions of a chessboard that were relevant or irrelevant to the best move in previous studies, we introduced n-mate tasks and sequentially manipulated their difficulty. Due to the complexity of the tasks, we monitored players' visual strategies in a fine-graded initial phase (different time intervals instead of analysing a fixed number of first fixations) of task-solving and for complete trials, employing non-obtrusive mobile eye tracking, multi-sensor observation and full-automatic annotation of decision-making. The results revealed significant expertise-dependent differences in visuocognitive performance based on a circumstantial spatial and temporal analysis. In order to provide more detailed results, for the first time the analyses were performed under the special consideration of different time intervals and spatial scalings. In summary, experts showed a significantly higher number of fixations on areas of interest and empty squares between pieces in the task processing than less-skilled players. However, they had a strikingly low total number of fixations on the whole board and in complete trials. As a conclusion, experts apply different visual search strategies in problem-solving. Moreover, experts' visuocognitive processing benefits from stored chunks of mating constellations.

Climate change-induced degradation of expert range maps drawn for kissing bugs (Hemiptera: Reduviidae) and long-standing current and future sampling gaps across the Americas.

Kissing bugs are the vectors of Trypanosoma cruzi, the etiological agent of Chagas disease (CD). Despite their epidemiological relevance, kissing bug species are under sampled in terms of their diversity and it is unclear what biases exist in available kissing bug data. Under climate change, range maps for kissing bugs may become less accurate as species shift their ranges to track climatic tolerance. Quantify inventory completeness in available kissing bug data. Assess how well range maps are at conveying information about current distributions and potential future distributions subject to shift under climate change. Intersect forecasted changes in kissing bug distributions with contemporary sampling gaps to identify regions for future sampling of the group. Identify whether a phylogenetic signal is present in expert range knowledge as more closely related species may be similarly well or lesser understood. We used species distribution models (SDM), specifically constructed from Bayesian additive regression trees, with Bioclim variables, to forecast kissing bug distributions into 2100 and intersect these with current sampling gaps to identify priority regions for sampling. Expert range maps were assessed by the agreement between the expert map and SDM generated occurrence probability. We used classical hypothesis testing methods as well as tests of phylogenetic signal to meet our objectives. Expert range maps vary in their quality of depicting current kissing bug distributions. Most expert range maps decline in their ability to convey information about kissing bug occurrence over time, especially in under sampled areas. We found limited evidence for a phylogenetic signal in expert range map performance. Expert range maps are not a perfect account of species distributions and may degrade in their ability to accurately convey distribution knowledge under future climates. We identify regions where future sampling of kissing bugs will be crucial for completing biodiversity inventories.

Designing the Framework of Specialization in the Recruitment of Non-Teaching Staff in Iran's Ministry of Education

Purpose: The present study aimed to design a framework for recruiting non-teaching educational staff in Iran’s Ministry of Education. Method: The present study pursues a qualitative approach. The statistical population of the research included experts and professors in the field of specialization and the senior managers of the General Department of Education of Semnan province in the academic year of 2023-2024, among which 15 were chosen as the samples using theoretical saturation sampling and targeted sampling. The research instruments were a semi-structured interview and a review of the existing documents on this subject. The collected data were validated using member checking. The data validity was evaluated and confirmed using the quality criterion (reliability). The research data analysis method was thematic analysis in three levels: basic themes, organizing themes and Global Themes. Findings: The data analysis led to the identification of 393 basic themes, 33 number-one organizing themes, and 13 number-2 organizing themes in the field of specialization in recruiting non-teaching educational staff in the Ministry of Education. These themes were categorized under 4 global themes including expertise management with two organizing themes (expertise management and performance management), resource management with four organizing themes (human, time, material, and information), change management with three organizing themes (change in organizational structure, change in competitive atmosphere, and change in attitude), and management of organizational development with four organizing themes (managerial capabilities, human resource planning, creation of specialization culture, and organizational interactions). Conclusion: The use of elite and expert employees in the nonteaching sector of the educational system, as a leading institution in providing extensive services to the general public, depends on the presence of a specialization pattern in the recruitment of non-teaching staff. Identification of specialization criteria in the non-teaching sector, which has been done in this research, helps with the formulation of an appropriate specialization model to employ skilled and specialist staff in this sector.

Accurate COVID-19 detection using full blood count data and machine learning

COVID-19 has spread rapidly worldwide in the past three years, triggering partial and full lockdowns globally. The successful control of the COVID-19 pandemic on a global scale depended heavily upon the accurate detection of COVID-19. However, the main diagnostic tests for COVID-19 have some significant limitations, e.g. the major nucleic acid (RT-PCR) tests while having a high sensitivity are time-consuming and require expensive equipment with the shortage of test kits in many countries. Antigen lateral flow tests have a lower sensitivity and they cannot be used during the early pandemic as well as usually more expensive than the full or complete blood count test used in this paper which can be potentially performed using a finger blood sample. The last decade has seen rapid growth of AI, particularly deep learning, which has found wide applications in medical image analysis, with results comparable to and even surpassing human expert performance. There have been several machine learning models reported for COVID-19 diagnostics or prognosis predictions, most of them based on CT and X-ray images. In this paper we have applied traditional machine learning and convolutional neural networks (CNNs) based deep learning to the blood test data obtained from hematology analyzers and demonstrated that the AI models can be used to detect COVID-19 with a high degree of accuracy (&gt;97%). The performance of different classifiers will be compared and discussed. The work should have potential applications in current COVID-19 and future pandemics.