A. B. M. Rezbaul Islam
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
A. B. M. Rezbaul Islam has a total of 20 co-authors. Here are the top co-authors by number of shared publications:
- Narasimha Shashidhar
Department of Computer Science, Sam Houston State University, Huntsville, TX 77341, USA
4 shared publications - Ahsan Ul Islam
Department of Computer Science, Sam Houston State University, Huntsville, TX 77340, USA
3 shared publications - Wadduwage Shanika Perera
Department of Computer Science, Sam Houston State University, Huntsville, TX 77340, USA
3 shared publications - Adebola Omogunwa
Department of Public Health, Sam Houston State University, Huntsville, TX, United States
2 shared publications - Amanda Scarbrough
Department of Public Health, Sam Houston State University, Huntsville, TX, United States
2 shared publications - Khalid M. Khan
Department of Public Health, College of Health Sciences, Sam Houston State University, Huntsville, USA
2 shared publications
Latest Publications
- Open Access
AI-Driven Digital Forensics: A Systematic Review of Capabilities, Limitations, and Research Challenges
The rapid growth of digital crime through cyber intrusions, ransomware, social engineering, and cloud exploitation has substantially increased the volume, heterogeneity, and complexity of digital evidence. Modern investigations span mobile devices, operating systems, volatile memory, networks, cloud infrastructures, and emerging environments such as the Internet of Things (IoT) and blockchain, challenging the scalability of conventional forensic workflows. Artificial Intelligence (AI) and Machine Learning (ML) offer promising capabilities to automate evidence analysis, identify complex patterns, and accelerate investigative processes. This systematic review provides a comprehensive and critical synthesis of AI/ML applications across major digital forensic domains, including file, memory, network, cloud, IoT, blockchain, mobile, social media, multimedia, biometric, vehicular, and drone forensics, as well as cross-cutting investigative tasks. A structured literature search and eligibility assessment identified 76 studies for qualitative synthesis. The review synthesizes the literature with attention to dataset availability, validation practices, reproducibility, explainability, adversarial robustness, and forensic admissibility. Beyond descriptive synthesis, this study introduces a six-dimensional task maturity framework based on observability, labelability, stationarity, time-to-insight, explainability, and adversarial robustness. An evidence audit links individual ratings to supporting references and qualifications. The framework assigns literature-informed ordinal scores using explicit dimensional criteria. Supporting references, rationales, and scope limitations are documented for all 96 task-dimension ratings. Sensitivity analysis examines the same profiles under alternative weights, thresholds, and individual rating changes. The reference mapping does not independently establish the validity of each numerical level. Contextual and interpretive judgment remains constrained by the need for human situational and legal reasoning. The review further identifies persistent gaps in benchmark datasets, real-world validation, explainability, adversarial testing, and legal defensibility. These findings motivate a research agenda centered on standardized evaluation protocols, robust and explainable forensic AI, improved benchmark resources, and accountable human–AI collaboration. Overall, this study provides a structured assessment of the current capabilities and limitations of AI/ML in digital forensics and identifies priorities for developing reliable, transparent, and legally defensible intelligent forensic systems.
Open AccessAI-Driven Digital Forensics: A Systematic Review of Capabilities, Limitations, and Research Challenges
The rapid growth of digital crime through cyber intrusions, ransomware, social engineering, and cloud exploitation has substantially increased the volume, heterogeneity, and complexity of digital evidence. Modern investigations span mobile devices, operating systems, volatile memory, networks, cloud infrastructures, and emerging environments such as the Internet of Things (IoT) and blockchain, challenging the scalability of conventional forensic workflows. Artificial Intelligence (AI) and Machine Learning (ML) offer promising capabilities to automate evidence analysis, identify complex patterns, and accelerate investigative processes. This systematic review provides a comprehensive and critical synthesis of AI/ML applications across major digital forensic domains, including file, memory, network, cloud, IoT, blockchain, mobile, social media, multimedia, biometric, vehicular, and drone forensics, as well as cross-cutting investigative tasks. A structured literature search and eligibility assessment identified 76 studies for qualitative synthesis. The review synthesizes the literature with attention to dataset availability, validation practices, reproducibility, explainability, adversarial robustness, and forensic admissibility. Beyond descriptive synthesis, this study introduces a six-dimensional task maturity framework based on observability, labelability, stationarity, time-to-insight, explainability, and adversarial robustness. An evidence audit links individual ratings to supporting references and qualifications. The framework assigns literature-informed ordinal scores using explicit dimensional criteria. Supporting references, rationales, and scope limitations are documented for all 96 task-dimension ratings. Sensitivity analysis examines the same profiles under alternative weights, thresholds, and individual rating changes. The reference mapping does not independently establish the validity of each numerical level. Contextual and interpretive judgment remains constrained by the need for human situational and legal reasoning. The review further identifies persistent gaps in benchmark datasets, real-world validation, explainability, adversarial testing, and legal defensibility. These findings motivate a research agenda centered on standardized evaluation protocols, robust and explainable forensic AI, improved benchmark resources, and accountable human–AI collaboration. Overall, this study provides a structured assessment of the current capabilities and limitations of AI/ML in digital forensics and identifies priorities for developing reliable, transparent, and legally defensible intelligent forensic systems.
- Open Access
- 21 September 2026
- Published by MDPI AG in Preprints.org
AI-Driven Digital Forensics: Capabilities, Limitations, and Research Challenges
The rapid growth of digital crime through cyber intrusions, ransomware, social engineering, and cloud exploitation has substantially increased the volume, heterogeneity, and complexity of digital evidence. Modern investigations span mobile devices, operating systems, volatile memory, networks, cloud infrastructures, and emerging environments such as the Internet of Things (IoT) and blockchain, challenging the scalability of conventional forensic workflows. Artificial Intelligence (AI) and Machine Learning (ML) offer promising capabilities to automate evidence analysis, identify complex patterns, and accelerate investigative processes. This survey provides a comprehensive and critical synthesis of AI/ML applications across major digital forensic domains, including file, memory, network, cloud, IoT, blockchain, mobile, social media, multimedia, biometric, vehicular, and drone forensics, as well as cross-cutting investigative tasks. The reviewed literature is synthesized with attention to dataset availability, validation practices, reproducibility, explainability, adversarial robustness, and forensic admissibility. Beyond descriptive synthesis, this study develops and applies a six-dimensional solvability framework that evaluates representative forensic tasks according to observability, labelability, stationarity, time-to-insight, explainability, and adversarial robustness. The framework classifies tasks as solvable, partially solvable, or unsolved under current technical and evidentiary conditions. The analysis identifies file-type identification and carving, multimedia manipulation detection, and biometric identification as comparatively solvable. At the same time, most surveyed tasks remain partially solvable because of dataset scarcity, limited cross-domain generalization, model opacity, and adversarial vulnerabilities. Contextual and interpretive judgment remains critically constrained because reliable forensic conclusions require human situational and legal reasoning. The survey further identifies persistent gaps in benchmark datasets, real-world validation, explainability, adversarial testing, and legal defensibility. These findings motivate a research agenda centered on standardized evaluation protocols, robust and explainable forensic AI, improved benchmark resources, and accountable human–AI collaboration. Overall, this study provides a structured assessment of the current capabilities and limitations of AI/ML in digital forensics and identifies priorities for developing reliable, transparent, and legally defensible intelligent forensic systems.
- 21 September 2026
- Published by MDPI AG in Preprints.org
Open AccessAI-Driven Digital Forensics: Capabilities, Limitations, and Research Challenges
The rapid growth of digital crime through cyber intrusions, ransomware, social engineering, and cloud exploitation has substantially increased the volume, heterogeneity, and complexity of digital evidence. Modern investigations span mobile devices, operating systems, volatile memory, networks, cloud infrastructures, and emerging environments such as the Internet of Things (IoT) and blockchain, challenging the scalability of conventional forensic workflows. Artificial Intelligence (AI) and Machine Learning (ML) offer promising capabilities to automate evidence analysis, identify complex patterns, and accelerate investigative processes. This survey provides a comprehensive and critical synthesis of AI/ML applications across major digital forensic domains, including file, memory, network, cloud, IoT, blockchain, mobile, social media, multimedia, biometric, vehicular, and drone forensics, as well as cross-cutting investigative tasks. The reviewed literature is synthesized with attention to dataset availability, validation practices, reproducibility, explainability, adversarial robustness, and forensic admissibility. Beyond descriptive synthesis, this study develops and applies a six-dimensional solvability framework that evaluates representative forensic tasks according to observability, labelability, stationarity, time-to-insight, explainability, and adversarial robustness. The framework classifies tasks as solvable, partially solvable, or unsolved under current technical and evidentiary conditions. The analysis identifies file-type identification and carving, multimedia manipulation detection, and biometric identification as comparatively solvable. At the same time, most surveyed tasks remain partially solvable because of dataset scarcity, limited cross-domain generalization, model opacity, and adversarial vulnerabilities. Contextual and interpretive judgment remains critically constrained because reliable forensic conclusions require human situational and legal reasoning. The survey further identifies persistent gaps in benchmark datasets, real-world validation, explainability, adversarial testing, and legal defensibility. These findings motivate a research agenda centered on standardized evaluation protocols, robust and explainable forensic AI, improved benchmark resources, and accountable human–AI collaboration. Overall, this study provides a structured assessment of the current capabilities and limitations of AI/ML in digital forensics and identifies priorities for developing reliable, transparent, and legally defensible intelligent forensic systems.
- 8 September 2026
- Published by Springer Nature
- p. 34-51
Exposing Vulnerabilities in AI-Based Malware Detectors: Adversarial, Backdoor, and Poisoning Attacks
- 8 September 2026
- Published by Springer Nature
- p. 34-51
Exposing Vulnerabilities in AI-Based Malware Detectors: Adversarial, Backdoor, and Poisoning Attacks
- Open Access
- 18 December 2025
- Published by Springer Nature in Research Square
Gender and Ethnicity Bias in Deep Learning
People's opinions and actions in everyday life are increasingly influenced by artificial intelligence. However, representation in the design of these technologies has the potential to quietly undo decades of progress in gender and ethnicity. These biases threaten the strides toward equality in these areas, casting a shadow over our progress. The concerns surrounding gender and ethnicity biases have pervaded numerous fields, none more prominently than within artificial intelligence, especially in pre-trained deep learning models. These models, celebrated for their ability to extract knowledge from extensive datasets, have immense potential to revolutionize society and decision-making. However, they are not impervious to the biases embedded in the data upon which they are trained. It raises the troubling possibility of unintentionally perpetuating and amplifying social biases linked to gender or ethnicity.The issue of gender bias in pre-trained deep learning models has gained significant traction in recent times. As these models have become increasingly ubiquitous across various applications, it has become evident that they often perpetuate and exacerbate long-standing gender biases inherent in the training data. This paper embarks on a methodical and empirically rigorous exploration, delving into the nuanced landscape of gender and ethnicity bias within a diverse array of pre-trained deep learning models. Through meticulous scrutiny of these models' performance about gender and ethnicity-based predictions, we aim to unearth invaluable insights regarding the presence, intricacies, and magnitude of bias.This research paper offers a comprehensive and empirically grounded examination of gender and ethnicity bias within a diverse range of pre-trained deep-learning models. This investigation involves a meticulous analysis of how these models perform when making predictions related to gender and ethnicity. By scrutinizing their predictions, the aim is to unearth valuable insights into the presence, nuances, and extent of these biases within AI systems.Furthermore, this work introduces an innovative, holistic solution to mitigate gender and ethnicity bias. We present CNN models strategically crafted to address and rectify biases about gender and ethnicity effectively. This model represents a pioneering step towards combating bias on multiple fronts within AI systems.This research thus contributes significantly to the broader understanding of bias within AI technologies. Simultaneously addressing gender and ethnicity bias and proposing a practical remedy and the way for more equitable and unbiased advancements in artificial intelligence. Through rigorous analysis and innovative solutions, we seek to ensure that AI systems respect and uphold the principles of fairness, inclusively, and diversity, thereby fostering a more just technological landscape for all.
- 18 December 2025
- Published by Springer Nature in Research Square
Open AccessGender and Ethnicity Bias in Deep Learning
People's opinions and actions in everyday life are increasingly influenced by artificial intelligence. However, representation in the design of these technologies has the potential to quietly undo decades of progress in gender and ethnicity. These biases threaten the strides toward equality in these areas, casting a shadow over our progress. The concerns surrounding gender and ethnicity biases have pervaded numerous fields, none more prominently than within artificial intelligence, especially in pre-trained deep learning models. These models, celebrated for their ability to extract knowledge from extensive datasets, have immense potential to revolutionize society and decision-making. However, they are not impervious to the biases embedded in the data upon which they are trained. It raises the troubling possibility of unintentionally perpetuating and amplifying social biases linked to gender or ethnicity.The issue of gender bias in pre-trained deep learning models has gained significant traction in recent times. As these models have become increasingly ubiquitous across various applications, it has become evident that they often perpetuate and exacerbate long-standing gender biases inherent in the training data. This paper embarks on a methodical and empirically rigorous exploration, delving into the nuanced landscape of gender and ethnicity bias within a diverse array of pre-trained deep learning models. Through meticulous scrutiny of these models' performance about gender and ethnicity-based predictions, we aim to unearth invaluable insights regarding the presence, intricacies, and magnitude of bias.This research paper offers a comprehensive and empirically grounded examination of gender and ethnicity bias within a diverse range of pre-trained deep-learning models. This investigation involves a meticulous analysis of how these models perform when making predictions related to gender and ethnicity. By scrutinizing their predictions, the aim is to unearth valuable insights into the presence, nuances, and extent of these biases within AI systems.Furthermore, this work introduces an innovative, holistic solution to mitigate gender and ethnicity bias. We present CNN models strategically crafted to address and rectify biases about gender and ethnicity effectively. This model represents a pioneering step towards combating bias on multiple fronts within AI systems.This research thus contributes significantly to the broader understanding of bias within AI technologies. Simultaneously addressing gender and ethnicity bias and proposing a practical remedy and the way for more equitable and unbiased advancements in artificial intelligence. Through rigorous analysis and innovative solutions, we seek to ensure that AI systems respect and uphold the principles of fairness, inclusively, and diversity, thereby fostering a more just technological landscape for all.
- Open Access
- 21 June 2024
- Published by MDPI AG in Electronics
- Vol. 13 (13) , 2438
Interpretability and Transparency of Machine Learning in File Fragment Analysis with Explainable Artificial Intelligence
Machine learning models are increasingly being used across diverse fields, including file fragment classification. As these models become more prevalent, it is crucial to understand and interpret their decision-making processes to ensure accountability, transparency, and trust. This research investigates the interpretability of four machine learning models used for file fragment classification through the lens of Explainable Artificial Intelligence (XAI) techniques. Specifically, we employ two prominent XAI methods, Shapley Additive Explanations (SHAP) and Local Interpretable Model-Agnostic Explanations (LIME), to shed light on the black-box nature of four machine learning models used for file fragment classification. By conducting a detailed analysis of the SHAP and LIME explanations, we demonstrate the effectiveness of these techniques in improving the interpretability of the models’ decision-making processes. Our analysis reveals that these XAI techniques effectively identify key features influencing each model’s predictions. The results also showed features that were critical to predicting specific classes. The ability to interpret and validate the decisions made by machine learning models in file fragment classification can enhance trust in these models and inform improvements for better accuracy and reliability. Our research highlights the importance of XAI techniques in promoting transparency and accountability in the application of machine learning models across diverse domains.
- 21 June 2024
- Published by MDPI AG in Electronics
- Vol. 13 (13) , 2438
Open AccessInterpretability and Transparency of Machine Learning in File Fragment Analysis with Explainable Artificial Intelligence
Machine learning models are increasingly being used across diverse fields, including file fragment classification. As these models become more prevalent, it is crucial to understand and interpret their decision-making processes to ensure accountability, transparency, and trust. This research investigates the interpretability of four machine learning models used for file fragment classification through the lens of Explainable Artificial Intelligence (XAI) techniques. Specifically, we employ two prominent XAI methods, Shapley Additive Explanations (SHAP) and Local Interpretable Model-Agnostic Explanations (LIME), to shed light on the black-box nature of four machine learning models used for file fragment classification. By conducting a detailed analysis of the SHAP and LIME explanations, we demonstrate the effectiveness of these techniques in improving the interpretability of the models’ decision-making processes. Our analysis reveals that these XAI techniques effectively identify key features influencing each model’s predictions. The results also showed features that were critical to predicting specific classes. The ability to interpret and validate the decisions made by machine learning models in file fragment classification can enhance trust in these models and inform improvements for better accuracy and reliability. Our research highlights the importance of XAI techniques in promoting transparency and accountability in the application of machine learning models across diverse domains.
- 23 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 25-30
Cotton Care: Harnessing Transfer Learning for Real-Time Health Assessment with Mobile Devices
Cotton frequently experiences substantial yield reduction due to disease afflictions. This paper introduces a proof-of-concept mobile application that employs transfer learning with pre-trained MobileNetV3 and NASNetMobile models to detect cotton plant diseases. The system addresses a four-class problem, identifying the health status of cotton plants or leaves. Using TensorFlow, these models are refined with publicly available data and deployed on an iOS app. This study also provides a performance assessment of both models, previously unexplored in this specific context. The methodology, showcasing significant potential for real-world implementations, recorded overall accuracies of 97.7% for NASNetMobile and 96.7% for MobileNetV3.
- 23 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 25-30
Cotton Care: Harnessing Transfer Learning for Real-Time Health Assessment with Mobile Devices
Cotton frequently experiences substantial yield reduction due to disease afflictions. This paper introduces a proof-of-concept mobile application that employs transfer learning with pre-trained MobileNetV3 and NASNetMobile models to detect cotton plant diseases. The system addresses a four-class problem, identifying the health status of cotton plants or leaves. Using TensorFlow, these models are refined with publicly available data and deployed on an iOS app. This study also provides a performance assessment of both models, previously unexplored in this specific context. The methodology, showcasing significant potential for real-world implementations, recorded overall accuracies of 97.7% for NASNetMobile and 96.7% for MobileNetV3.
- Open Access
- 30 May 2023
- Published by JMIR Publications Inc. in JMIR Formative Research
- Vol. 7 (1) , e45434
An Artificial Intelligence–Based Smartphone App for Assessing the Risk of Opioid Misuse in Working Populations Using Synthetic Data: Pilot Development Study
Background: Opioid use disorder (OUD) is an addiction crisis in the United States. As recent as 2019, more than 10 million people have misused or abused prescription opioids, making OUD one of the leading causes of accidental death in the United States. Workforces that are physically demanding and laborious in the transportation, construction and extraction, and health care industries are prime targets for OUD due to high-risk occupational activities. Because of this high prevalence of OUD among working populations in the United States, elevated workers’ compensation and health insurance costs, absenteeism, and declined productivity in workplaces have been reported. Objective: With the emergence of new smartphone technologies, health interventions can be widely used outside clinical settings via mobile health tools. The major objective of our pilot study was to develop a smartphone app that can track work-related risk factors leading to OUD with a specific focus on high-risk occupational groups. We used synthetic data analyzed by applying a machine learning algorithm to accomplish our objective. Methods: To make the OUD assessment process more convenient and to motivate potential patients with OUD, we developed a smartphone-based app through a step-by-step process. First, an extensive literature survey was conducted to list a set of critical risk assessment questions that can capture high-risk behaviors leading to OUD. Next, a review panel short-listed 15 questions after careful evaluation with specific emphasis on physically demanding workforces—9 questions had two, 5 questions had five, and 1 question had three response options. Instead of human participant data, synthetic data were used as user responses. Finally, an artificial intelligence algorithm, naive Bayes, was used to predict the OUD risk, trained with the synthetic data collected. Results: The smartphone app we have developed is functional as tested with synthetic data. Using the naive Bayes algorithm on collected synthetic data, we successfully predicted the risk of OUD. This would eventually create a platform to test the functionality of the app further using human participant data. Conclusions: The use of mobile health techniques, such as our mobile app, is highly promising in predicting and offering mitigation plans for disease detection and prevention. Using a naive Bayes algorithm model along with a representational state transfer (REST) application programming interface and cloud-based data encryption storage, respondents can guarantee their privacy and accuracy in estimating their risk. Our app offers a tailored mitigation strategy for specific workforces (eg, transportation and health care workers) that are most impacted by OUD. Despite the limitations of the study, we have developed a robust methodology and believe that our app has the potential to help reduce the opioid crisis.
- 30 May 2023
- Published by JMIR Publications Inc. in JMIR Formative Research
- Vol. 7 (1) , e45434
Open AccessAn Artificial Intelligence–Based Smartphone App for Assessing the Risk of Opioid Misuse in Working Populations Using Synthetic Data: Pilot Development Study
Background: Opioid use disorder (OUD) is an addiction crisis in the United States. As recent as 2019, more than 10 million people have misused or abused prescription opioids, making OUD one of the leading causes of accidental death in the United States. Workforces that are physically demanding and laborious in the transportation, construction and extraction, and health care industries are prime targets for OUD due to high-risk occupational activities. Because of this high prevalence of OUD among working populations in the United States, elevated workers’ compensation and health insurance costs, absenteeism, and declined productivity in workplaces have been reported. Objective: With the emergence of new smartphone technologies, health interventions can be widely used outside clinical settings via mobile health tools. The major objective of our pilot study was to develop a smartphone app that can track work-related risk factors leading to OUD with a specific focus on high-risk occupational groups. We used synthetic data analyzed by applying a machine learning algorithm to accomplish our objective. Methods: To make the OUD assessment process more convenient and to motivate potential patients with OUD, we developed a smartphone-based app through a step-by-step process. First, an extensive literature survey was conducted to list a set of critical risk assessment questions that can capture high-risk behaviors leading to OUD. Next, a review panel short-listed 15 questions after careful evaluation with specific emphasis on physically demanding workforces—9 questions had two, 5 questions had five, and 1 question had three response options. Instead of human participant data, synthetic data were used as user responses. Finally, an artificial intelligence algorithm, naive Bayes, was used to predict the OUD risk, trained with the synthetic data collected. Results: The smartphone app we have developed is functional as tested with synthetic data. Using the naive Bayes algorithm on collected synthetic data, we successfully predicted the risk of OUD. This would eventually create a platform to test the functionality of the app further using human participant data. Conclusions: The use of mobile health techniques, such as our mobile app, is highly promising in predicting and offering mitigation plans for disease detection and prevention. Using a naive Bayes algorithm model along with a representational state transfer (REST) application programming interface and cloud-based data encryption storage, respondents can guarantee their privacy and accuracy in estimating their risk. Our app offers a tailored mitigation strategy for specific workforces (eg, transportation and health care workers) that are most impacted by OUD. Despite the limitations of the study, we have developed a robust methodology and believe that our app has the potential to help reduce the opioid crisis.
- 27 May 2023
- Published by Springer Nature
- p. 316-328
Statistical Approaches to File Fragment Analysis
- Open Access
- 31 December 2022
- Published by JMIR Publications Inc. in JMIR Preprints
An Artificial Intelligence–Based Smartphone App for Assessing the Risk of Opioid Misuse in Working Populations Using Synthetic Data: Pilot Development Study (Preprint)
BACKGROUND Opioid use disorder (OUD) is an addiction crisis in the United States. As recent as 2019, more than 10 million people have misused or abused prescription opioids, making OUD one of the leading causes of accidental death in the United States. Workforces that are physically demanding and laborious in the transportation, construction and extraction, and health care industries are prime targets for OUD due to high-risk occupational activities. Because of this high prevalence of OUD among working populations in the United States, elevated workers’ compensation and health insurance costs, absenteeism, and declined productivity in workplaces have been reported. OBJECTIVE With the emergence of new smartphone technologies, health interventions can be widely used outside clinical settings via mobile health tools. The major objective of our pilot study was to develop a smartphone app that can track work-related risk factors leading to OUD with a specific focus on high-risk occupational groups. We used synthetic data analyzed by applying a machine learning algorithm to accomplish our objective. METHODS To make the OUD assessment process more convenient and to motivate potential patients with OUD, we developed a smartphone-based app through a step-by-step process. First, an extensive literature survey was conducted to list a set of critical risk assessment questions that can capture high-risk behaviors leading to OUD. Next, a review panel short-listed 15 questions after careful evaluation with specific emphasis on physically demanding workforces—9 questions had two, 5 questions had five, and 1 question had three response options. Instead of human participant data, synthetic data were used as user responses. Finally, an artificial intelligence algorithm, naive Bayes, was used to predict the OUD risk, trained with the synthetic data collected. RESULTS The smartphone app we have developed is functional as tested with synthetic data. Using the naive Bayes algorithm on collected synthetic data, we successfully predicted the risk of OUD. This would eventually create a platform to test the functionality of the app further using human participant data. CONCLUSIONS The use of mobile health techniques, such as our mobile app, is highly promising in predicting and offering mitigation plans for disease detection and prevention. Using a naive Bayes algorithm model along with a representational state transfer (REST) application programming interface and cloud-based data encryption storage, respondents can guarantee their privacy and accuracy in estimating their risk. Our app offers a tailored mitigation strategy for specific workforces (eg, transportation and health care workers) that are most impacted by OUD. Despite the limitations of the study, we have developed a robust methodology and believe that our app has the potential to help reduce the opioid crisis.
- 31 December 2022
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessAn Artificial Intelligence–Based Smartphone App for Assessing the Risk of Opioid Misuse in Working Populations Using Synthetic Data: Pilot Development Study (Preprint)
BACKGROUND Opioid use disorder (OUD) is an addiction crisis in the United States. As recent as 2019, more than 10 million people have misused or abused prescription opioids, making OUD one of the leading causes of accidental death in the United States. Workforces that are physically demanding and laborious in the transportation, construction and extraction, and health care industries are prime targets for OUD due to high-risk occupational activities. Because of this high prevalence of OUD among working populations in the United States, elevated workers’ compensation and health insurance costs, absenteeism, and declined productivity in workplaces have been reported. OBJECTIVE With the emergence of new smartphone technologies, health interventions can be widely used outside clinical settings via mobile health tools. The major objective of our pilot study was to develop a smartphone app that can track work-related risk factors leading to OUD with a specific focus on high-risk occupational groups. We used synthetic data analyzed by applying a machine learning algorithm to accomplish our objective. METHODS To make the OUD assessment process more convenient and to motivate potential patients with OUD, we developed a smartphone-based app through a step-by-step process. First, an extensive literature survey was conducted to list a set of critical risk assessment questions that can capture high-risk behaviors leading to OUD. Next, a review panel short-listed 15 questions after careful evaluation with specific emphasis on physically demanding workforces—9 questions had two, 5 questions had five, and 1 question had three response options. Instead of human participant data, synthetic data were used as user responses. Finally, an artificial intelligence algorithm, naive Bayes, was used to predict the OUD risk, trained with the synthetic data collected. RESULTS The smartphone app we have developed is functional as tested with synthetic data. Using the naive Bayes algorithm on collected synthetic data, we successfully predicted the risk of OUD. This would eventually create a platform to test the functionality of the app further using human participant data. CONCLUSIONS The use of mobile health techniques, such as our mobile app, is highly promising in predicting and offering mitigation plans for disease detection and prevention. Using a naive Bayes algorithm model along with a representational state transfer (REST) application programming interface and cloud-based data encryption storage, respondents can guarantee their privacy and accuracy in estimating their risk. Our app offers a tailored mitigation strategy for specific workforces (eg, transportation and health care workers) that are most impacted by OUD. Despite the limitations of the study, we have developed a robust methodology and believe that our app has the potential to help reduce the opioid crisis.
- Open Access
- 14 December 2021
- Published by Macrothink Institute, Inc. in Journal of Education and Training
- Vol. 9 (2) , 40-50
Concept App Development for National Library of Virtual Manipulatives (NLVM): An Initiative
An NSF-supported project, the National Library of Virtual Manipulatives (NLVM), was started in 1999 to develop a library of entirely interactive and web-based virtual manipulatives or concept tutorials. It was mainly in Java applets for mathematics instruction (K-12 emphasis). The NLVM was a resource to develop mathematics classrooms from which teachers could independently draw upon. The materials were also significant in the mathematical training of both in-service and pre-service instructors. However, twenty years later, with the proliferation of touch-screen devices, this fantastic resource is in a crisis of becoming lost to future generations. It has been seen that many computers no longer run Java. Besides, the older version of Java used in the initial advancement of the NLVM is no longer supported. Hence, it is the right time to rethink the NLVM for current, widely available touch screen devices. This is essential to bolster the offerings for both prekindergarten, K- 12, and university settings. Keeping the situation in mind, we have constructed four virtual manipulative applications compatible with touch-screen devices based on the outdated java-script types of virtual manipulatives available through the NLVM. The four concept apps are being conducted with children in classroom settings starting in Spring 2021. It will allow for further revisions of the apps if necessary. Then, the apps from this project would be combined as part of a more significant, federally, or privately funded external grant proposal to re-imagine and broaden upon all the virtual manipulatives currently found in the NLVM.
- 14 December 2021
- Published by Macrothink Institute, Inc. in Journal of Education and Training
- Vol. 9 (2) , 40-50
Open AccessConcept App Development for National Library of Virtual Manipulatives (NLVM): An Initiative
An NSF-supported project, the National Library of Virtual Manipulatives (NLVM), was started in 1999 to develop a library of entirely interactive and web-based virtual manipulatives or concept tutorials. It was mainly in Java applets for mathematics instruction (K-12 emphasis). The NLVM was a resource to develop mathematics classrooms from which teachers could independently draw upon. The materials were also significant in the mathematical training of both in-service and pre-service instructors. However, twenty years later, with the proliferation of touch-screen devices, this fantastic resource is in a crisis of becoming lost to future generations. It has been seen that many computers no longer run Java. Besides, the older version of Java used in the initial advancement of the NLVM is no longer supported. Hence, it is the right time to rethink the NLVM for current, widely available touch screen devices. This is essential to bolster the offerings for both prekindergarten, K- 12, and university settings. Keeping the situation in mind, we have constructed four virtual manipulative applications compatible with touch-screen devices based on the outdated java-script types of virtual manipulatives available through the NLVM. The four concept apps are being conducted with children in classroom settings starting in Spring 2021. It will allow for further revisions of the apps if necessary. Then, the apps from this project would be combined as part of a more significant, federally, or privately funded external grant proposal to re-imagine and broaden upon all the virtual manipulatives currently found in the NLVM.