A Survey on Deep Learning Pouyanfar, Samira; Sadiq, Saad; Yan, Yilin ...
ACM computing surveys,
09/2019, Letnik:
51, Številka:
5
Journal Article
Recenzirano
The field of machine learning is witnessing its golden era as deep learning slowly becomes the leader in this domain. Deep learning uses multiple layers to represent the abstractions of data to build ...computational models. Some key enabler deep learning algorithms such as generative adversarial networks, convolutional neural networks, and model transfers have completely changed our perception of information processing. However, there exists an aperture of understanding behind this tremendously fast-paced domain, because it was never previously represented from a multiscope perspective. The lack of core understanding renders these powerful methods as black-box machines that inhibit development at a fundamental level. Moreover, deep learning has repeatedly been perceived as a silver bullet to all stumbling blocks in machine learning, which is far from the truth. This article presents a comprehensive review of historical and recent state-of-the-art approaches in visual, audio, and text processing; social network analysis; and natural language processing, followed by the in-depth analysis on pivoting and groundbreaking advances in deep learning applications. It was also undertaken to review the issues faced in deep learning such as unsupervised learning, black-box models, and online learning and to illustrate how these challenges can be transformed into prolific future research avenues.
•Conducts a systematic quantitative literature review of industry 4.0 literature.•Develops a holistic framework based on the most recurrent research themes.•Finds seven communities and three research ...clusters using SNA.•Industry 4.0 in services industries is neglected.•Combination of theoretical approaches is necessary to deal with digitized services.
The “industry 4.0" phenomenon is expected to influence almost every aspect of business value chains, and hence it has been increasingly analyzed by management scholars. However, the overarching intellectual structure emerging from this new stream of literature has not yet been synthesized in a framework nor critically discussed. Furthermore, despite being part of the rhetoric in several recent industrial governmental plans, industry 4.0 in service sectors has not been systematically reviewed to date. By leveraging a systematic quantitative literature review, a data-driven approach and a quantitative methodology—embedding both bibliographic coupling and network analysis techniques—this study provides a clear visualization of the emerging intellectual structure of industry 4.0 in management studies. We also develop a framework based on the most recurrent themes emerging from the results of bibliometric and network analyses—the latter could be used by management scholars to understand studies surrounding industry 4.0. As service businesses can create and capture value generated through the 4th Industrial Revolution as well as manufacturing firms, we suggest that scholarly attention should also be directed toward the service industries and provide a research agenda.
In consensus-based multiple attribute group decision making (MAGDM) problems, it is frequent that some experts exhibit non-cooperative behaviors owing to the different areas to which they may belong ...and the different (sometimes conflicting) interests they might present. This may adversely affect the overall efficiency of the consensus reaching process, especially when some uncooperative behaviors by experts arise. To this end, this paper develops a novel consensus framework based on social network analysis (SNA) to deal with non-cooperative behaviors. In the proposed SNA-based consensus framework, a trust propagation and aggregation mechanism to yield experts’ weights from the social trust network is presented, and the obtained weights of experts are then integrated into the consensus-based MAGDM framework. Meanwhile, a non-cooperative behavior analysis module is designed to analyze the behaviors of experts. Based on the results of such analysis during the consensus process, each expert can express and modify the trust values pertaining other experts in the social trust network. As a result, both the social trust network and the weights of experts derived from it are dynamically updated in parallel. A simulation and comparison study is presented to demonstrate the efficiency of the SNA-based consensus framework for coping with non-cooperative behaviors.
Statistical network models describing multivariate dependency structures in psychological data have gained increasing popularity. Such comparably novel statistical techniques require specific ...guidelines to make them accessible to the research community. So far, researchers have provided tutorials guiding the estimation of networks and their accuracy. However, there is currently little guidance in determining what parts of the analyses and results should be documented in a scientific report. A lack of such reporting standards may foster researcher degrees of freedom and could provide fertile ground for questionable reporting practices. Here, we introduce reporting standards for network analyses in cross-sectional data, along with a tutorial and two examples. The presented guidelines are aimed at researchers as well as the broader scientific community, such as reviewers and journal editors evaluating scientific work. We conclude by discussing how the network literature specifically can benefit from such guidelines for reporting and transparency.
Translational Abstract
In recent years, network models have become increasingly popular in the field of psychology. Such comparably novel statistical techniques require specific guidelines to make them accessible to the research community. So far, researchers have provided tutorials guiding how network analysis can be applied to psychological data. However, there is currently little guidance in determining what parts of the analyses and results should be documented in a scientific report. A lack of such reporting standards may result in researchers being confronted with too much choice in reporting their results, which in turn might provide fertile ground for questionable reporting practices. Here, we introduce reporting standards for network analyses in cross-sectional data, along with a tutorial and two examples. The presented guidelines are aimed at researchers as well as the broader scientific community, such as reviewers and journal editors evaluating scientific work. We conclude by discussing how the network literature specifically can benefit from such guidelines for reporting and transparency.
As part of business and management studies, research works addressed blockchain technology (BCT) in logistics and supply chain management (LSCM) first in 2016. Increasing levels of interest from ...researchers and practitioners alike have led to an increasing number of studies from both ends; however, a thorough bibliometric- and cocitation network analysis of BCT in LSCM research has not been carried out so far. To address this gap and to build a basis for future research endeavors, this article provides a bibliometric analysis on BCT, comprising data from 613 articles from academic supply chain research. It is therefore an easy-to-access entry point for academics and practitioners into the topic of BCT in LSCM. This study aims to understand the status of research of BCT in LSCM. To present the results, this article employs a bibliometric analysis methodology. It adopts a citation network analysis and a cocitation analysis. Based on a cocitation analysis, this article classifies the existing literature into five different research clusters, including theoretical sensemaking, conceptualizing and testing blockchain applications, framing BCT into supply chains, the technical design of BCT applications for real-world LSCM applications, and the role of BCT within digital supply chains.
Depression ranks as one of the top five contributors to ill health in youth, the most formative period in life. Extensive research has highlighted the significant role of impulsivity in understanding ...depression. However, there has been limited exploration into how each dimension of impulsivity uniquely affect depressive symptoms, especially across crucial developmental stages like adolescence and young adulthood.
This study investigates the unique relationships between impulsivity (assessed by the short UPPS-P scale) and depression (assessed by the Patient Health Questionnaire-9) via network analysis. We analysed data from a total of 2296 participants, comprising 858 adolescents aged 14–17 years and 1438 young adults aged 18–25 years, to estimate both a combined network and age-group specific networks. Key features of the networks, including their structure, global connectivity, and bridge nodes, were compared.
The results indicated that age differentially impacts individual depression symptoms, both directly and indirectly, via impulsivity dimensions. The comparison test revealed consistent network structures between the two age groups, with several robust pathways, such as lack of perseverance to concentration difficulties, sensation seeking to suicidal ideation, and negative urgency to feelings of worthlessness. Negative urgency and lack of perseverance were identified as bridge nodes across the two networks.
The study employed a cross-sectional design, which limits the ability to estimate causal or temporal relationships.
The current findings highlight the significance of tailoring intervention strategies to individual symptom profiles and assessing negative urgency and lack of perseverance as potential early targets for depression among youth.
•Examined relationships between impulsivity and depression in adolescents and young adults via network analysis.•Age differentially impacts individual depression symptoms, both directly and indirectly, via impulsivity dimensions.•Symptomatic pathways between impulsivity dimensions and depression symptoms are consistent across age groups.•The overall connectivity of networks varies among different age groups.•Negative urgency and lack of perseverance consistently emerged as bridge nodes across age groups.
Exploring networks of mental and behavioral problems in children and adolescents may identify differences between one-child and multi-child families. This study compared the network structures of ...mental and behavioral problems in children and adolescents in one-child families versus multi-child families based on a nationwide survey.
Propensity score matching (PSM) was used to match children and adolescents from one-child families with those from multi-child families. Mental and behavioral problems were assessed using the Achenbach's Child Behavior Checklist (CBCL) with eight syndromal subscales. In the network analysis, strength centrality index was used to estimate central symptoms, and case-dropping bootstrap method was used to assess network stability.
The study included 39,648 children and adolescents (19,824 from one-child families and 19,824 from multi-child families). Children and adolescents from multi-child families exhibited different network structure and higher global strength compared to those from one-child families. In one-child families, the most central symptoms were “Social problems”, “Anxious/depressed” and “Withdrawn/depressed”, while in multi-child families, the most central symptoms were “Social problems”, “Rule-breaking behavior” and “Anxious/depressed”.
Differences in mental and behavioral problems among children and adolescents between one-child and multi-child families were found. To address these problems, interventions targeting “Social problems” and “Anxious/depressed” symptoms should be developed for children and adolescents in both one-child and multi-child families, while other interventions targeting “Withdrawn/depressed” and “Rule-breaking behavior” symptoms could be useful for those in one-child and multi-child families, respectively.
•This study was based on a nationwide survey, using propensity score matching for balancing baseline characteristics.•Children and adolescents from multi-child families were more inclined to externalize psychological problems.•Children and adolescents from one-child families were more prone to internalize psychological problems.•There are obvious gender and age differences in network of mental and behavioral problems in children and adolescents.
This study compares the interaction patterns of a novice and an experienced instructor using Social Network Analysis (SNA) and content analysis and explores how students' interactions, degrees of ...satisfaction, and cognitive presence differ according to the different interaction patterns of the two instructors. Results showed some differences in the interaction characteristics between the sections. First, the experienced instructor was the most powerful actor in the course, while some students in the novice instructor's section showed higher outdegree centrality than the instructor. In addition, the novice instructor's section was a more active network than the experienced instructor's section in which the instructor showed the highest outdegree and indegree and also seemed to have more reciprocal relations. In terms of satisfaction and cognitive presence levels, the students in the experienced instructor's section in which the instructor focused more on triggering events or exploration activities, reported higher satisfaction than the students in the novice instructor's section. However, there was no significant difference in students' cognitive presence levels. A key finding of research suggests that instructors need to balance their participation, stimulate students' curiosity, and encourage brainstorming-rather than directly offering solutions-to improve students' satisfaction in asynchronous discussion-based online learning. This research also indicates that well-designed discussion topics may contribute more to developing students' cognitive presence than the instructor's interaction patterns. Finally, this research highlights the effectiveness of SNA and content analysis to explore instructors' and students' interactions on discussion boards.
Während qualitative Ansätze in der sozialen Netzwerkanalyse florieren, sind Forschungsprozesse und insbesondere die Datenanalyse zumeist von einem strukturalen netzwerkanalytischen Paradigma geprägt. ...Zudem existieren unzureichend qualitativ-interpretative Ansätze zur Untersuchung sozialer Netzwerkdaten. Um diese Forschungslücke zu schließen, entwerfen und explizieren wir ein qualitatives Analyseverfahren, das auf dem Cultural Turn der sozialen Netzwerkanalyse aufbaut und sowohl subjektive Deutungsmuster als auch historisch/prozessuale Konfigurationen erfassen soll. Wir formulieren eine biografische netzwerkanalytische Perspektive, in der wir die Entwicklung eines sozialen Netzwerkes in der Lebensgeschichte analysieren. Am Beispiel einer Fallstudie aus einem Forschungsprojekt zu transnationalen sozialen Unterstützungsnetzwerken älterer Migrant*innen in Perth explizieren wir das Verfahren der biografisch rekonstruktiven Netzwerkanalyse (BRNA). BRNA ist ein kooperativ entwickeltes analytisches Verfahren der Erhebung und der Auswertung sozialer Netzwerkdaten. Bei der BRNA-Datenerhebung werden das narrativ-biografische Interview und ego-zentrische Netzwerkkarten trianguliert. Bei der Datenanalyse folgen wir biografisch-rekonstruktiven Forschungsprinzipien und Verfahren, um die Dynamiken sozialer Netzwerke in der Lebensgeschichte zu rekonstruieren und nachzuvollziehen.
•Substance use and abuse has substantially increased during the COVID-19 pandemic.•Network analyses were conducted to understand COVID-19-related substance abuse.•Substance abuse was related to ...COVID-19-related traumatic stress symptoms.•Substance abuse was also related to noncompliance with social distancing.•Findings have potentially importance clinical and public health implications.
Research shows that there has been a substantial increase in substance use and abuse during the COVID-19 pandemic, and that substance use/abuse is a commonly reported way of coping with anxiety concerning COVID-19. Anxiety about COVID-19 is more than simply worry about infection. Research provides evidence of a COVID Stress Syndrome characterized by (1) worry about the dangers of COVID-19 and worry about coming into contact with coronavirus contaminated objects or surfaces, (2) worry about the personal socioeconomic impact of COVID-19, (3) xenophobic worries that foreigners are spreading COVID-19, (4) COVID-19-related traumatic stress symptoms (e.g., nightmares), and (5) COVID-19-related compulsive checking and reassurance-seeking. These form a network of interrelated nodes. Research also provides evidence of another constellation or “syndrome”, characterized by (1) belief that one has robust physical health against COVID-19, (2) belief that the threat of COVID-19 has been exaggerated, and (3) disregard for social distancing. These also form a network of nodes known as a COVID-19 Disregard Syndrome. The present study, based on a population-representative sample of 3075 American and Canadian adults, sought to investigate how these syndromes are related to substance use and abuse. We found substantial COVID-19-related increases in alcohol and drug use. Network analyses indicated that although the two syndromes are negatively correlated with one another, they both have positive links to alcohol and drug abuse. More specifically, COVID-19-related traumatic stress symptoms and the tendency to disregard social distancing were both linked to substance abuse. Clinical and public health implications are discussed.