Papers by Aravind Sasidharan Pillai
International Research Journal of Modernization in Engineering Technology and Science, 2025
Integrating deep learning into medical imaging has significantly advanced diagnostic accuracy, pa... more Integrating deep learning into medical imaging has significantly advanced diagnostic accuracy, particularly in classifying X-ray images for pulmonary conditions. This paper presents a comprehensive review of deep learning methodologies applied to X-ray classification, emphasizing their effectiveness in diagnosing diseases such as pneumonia, tuberculosis, and COVID-19. We explore various deep learning architectures, including Convolutional Neural Networks (CNNs), ResNet, VGG, and EfficientNet, highlighting their comparative performance in X-ray analysis. The study also examines key challenges, including dataset quality, class imbalance, and model interpretability, which impact the clinical applicability of these technologies. Furthermore, ethical considerations such as data privacy, algorithmic bias, and transparency in decisionmaking are discussed. By synthesizing current literature, we provide insights into the ongoing advancements in deep learning for medical imaging and outline future research directions to enhance model robustness and integration into clinical practice. This review aims to contribute to the evolving landscape of AI-driven medical diagnostics, underscoring the transformative potential of deep learning in radiology.
The integration of artificial intelligence (AI) in healthcare has revolutionized the prediction a... more The integration of artificial intelligence (AI) in healthcare has revolutionized the prediction and management of cardiac diseases. Given that cardiovascular diseases (CVDs) are a leading cause of mortality worldwide, there is an urgent need for innovative diagnostic methodologies. This paper explores the intersection of AI and cardiac disease prediction, emphasizing machine learning and deep learning applications. By synthesizing existing literature, this study critically evaluates the advantages of AI over traditional diagnostic methods, highlighting improvements in predictive accuracy and personalized treatment strategies. Furthermore, this review addresses methodological inconsistencies and ethical concerns regarding AI deployment in clinical settings. The findings underscore the potential of AI-driven models to enhance early detection and risk assessment in cardiovascular healthcare. However, challenges such as data biases, lack of transparency, and the need for regulatory frameworks must be addressed to harness AI's potential in cardiology fully. This study contributes to the growing discourse on AI in medical diagnostics, offering insights for future research and clinical implementation.
Journal of Emerging Technologies and Innovative Research, 2024
Traffic congestion remains a persistent issue in urban areas, leading to increased travel time, f... more Traffic congestion remains a persistent issue in urban areas, leading to increased travel time, fuel consumption, and environmental pollution. Traditional traffic management systems often fall short in dynamically adapting to real-time conditions. This research explores the implementation of Artificial Intelligence (AI) to optimize traffic flow and reduce congestion. By leveraging advanced AI techniques such as machine learning, neural networks, and computer vision, we develop predictive models for traffic management. These models are trained on extensive traffic data and tested in simulated environments to evaluate their effectiveness. The study also examines case studies from cities that have successfully integrated AI into their traffic systems, highlighting the benefits and challenges encountered. Our findings indicate that AI-driven traffic management significantly improves traffic flow, reduces congestion, and offers a scalable solution for modern urban planning. The study concludes with recommendations for policymakers and future research directions to enhance the implementation of AI in traffic management.
International Research Journal of Modernization in Engineering Technology and Science
Fake job postings have become prevalent in the online job market, posing significant challenges t... more Fake job postings have become prevalent in the online job market, posing significant challenges to job seekers and employers. Despite the growing need to address this problem, there is limited research that leverages deep learning techniques for the detection of fraudulent job advertisements. This study aims to fill the gap by employing a Bidirectional Long Short-Term Memory (Bi-LSTM) model to identify fake job advertisements. Our approach considers both numeric and text features, effectively capturing the underlying patterns and relationships within the data. The proposed model demonstrates a superior performance, achieving a 0.91 ROC AUC score and a 98.71% accuracy rate, indicating its potential for practical applications in the online job market. The findings of this research contribute to the development of robust, automated tools that can help combat the proliferation of fake job postings and improve the overall integrity of the job search process. Moreover, we discuss challenges, future research directions, and ethical considerations related to our approach, aiming to inspire further exploration and development of practical solutions to combat online job fraud.
Zenodo (CERN European Organization for Nuclear Research), Nov 20, 2022
Cardiac disease, which includes a variety of diseases that affect the heart, is a leading cause o... more Cardiac disease, which includes a variety of diseases that affect the heart, is a leading cause of death worldwide. One of every four deaths in the United States is due to heart disease. This means that approximately 610,000 people die from the disease each year. Heart disease is much easier to treat if detected early. Machine learning can play a crucial role in early detection and save lives. This research aims to develop an artificial intelligence-based system that identifies patients who are more likely to develop heart disease based on their medical history. The heart disease dataset from the UCI Machine Learning Repository was used for training and validation. Traditional classification techniques such as logistic regression, random forest, gradient boosting, and extreme gradient boosting were used as base models, and the results were compared with the Tabnet model. Tabnet is a new robust, interpretable, deeplearning architecture for tabular data. TabNet uses sequential attention to choose which features to conclude from at each decision step, focusing its learning ability on the most salient features, allowing for interpretability and more efficient learning. Promising results were obtained and validated using ROC curves, accuracy, precision, sensitivity, specificity, and confusion matrices. The Tabnet deep learning model outperformed the others, achieving 94% accuracy, ROC score of 0.94, and specificity and sensitivity greater than 0.93.
Journal of Intelligent Learning Systems and Applications
In this era of pandemic, the future of healthcare industry has never been more exciting. Artifici... more In this era of pandemic, the future of healthcare industry has never been more exciting. Artificial intelligence and machine learning (AI & ML) present opportunities to develop solutions that cater for very specific needs within the industry. Deep learning in healthcare had become incredibly powerful for supporting clinics and in transforming patient care in general. Deep learning is increasingly being applied for the detection of clinically important features in the images beyond what can be perceived by the naked human eye. Chest X-ray images are one of the most common clinical method for diagnosing a number of diseases such as pneumonia, lung cancer and many other abnormalities like lesions and fractures. Proper diagnosis of a disease from X-ray images is often challenging task for even expert radiologists and there is a growing need for computerized support systems due to the large amount of information encoded in X-Ray images. The goal of this paper is to develop a lightweight solution to detect 14 different chest conditions from an X ray image. Given an X-ray image as input, our classifier outputs a label vector indicating which of 14 disease classes does the image fall into. Along with the image features, we are also going to use non-image features available in the data such as X-ray view type, age, gender etc. The original study conducted Stanford ML Group is our base line. Original study focuses on predicting 5 diseases. Our aim is to improve upon previous work, expand prediction to 14 diseases and provide insight for future chest radiography research.
Journal of Emerging Technologies and Innovative Research, 2024
Traffic congestion remains a persistent issue in urban areas, leading to increased travel time, f... more Traffic congestion remains a persistent issue in urban areas, leading to increased travel time, fuel consumption, and environmental pollution. Traditional traffic management systems often fall short in dynamically adapting to real-time conditions. This research explores the implementation of Artificial Intelligence (AI) to optimize traffic flow and reduce congestion. By leveraging advanced AI techniques such as machine learning, neural networks, and computer vision, we develop predictive models for traffic management. These models are trained on extensive traffic data and tested in simulated environments to evaluate their effectiveness. The study also examines case studies from cities that have successfully integrated AI into their traffic systems, highlighting the benefits and challenges encountered. Our findings indicate that AI-driven traffic management significantly improves traffic flow, reduces congestion, and offers a scalable solution for modern urban planning. The study concludes with recommendations for policymakers and future research directions to enhance the implementation of AI in traffic management.
This research introduces an automated, Natural Language Processing (NLP)-based method for assembl... more This research introduces an automated, Natural Language Processing (NLP)-based method for assembling students into groups based on shared interests, extracted from personal narratives. For the experiment, each student in the class was required to compose several stories, ranging from 300 to 400 words, to facilitate the extraction of common phrases. These phrases were then used to cluster students according to shared interests revealed in their personal stories. The study applied the Rapid Automatic Keyword Extraction (RAKE) algorithm, an unsupervised and languageagnostic technique for extracting keywords. This method is distinguished by its independence from specific linguistic structures, rendering it broadly applicable across various types of documents and fields. The RAKE algorithm operates through several distinct phases: The first phase involves removing all stopwords and phrase delimiters from the text. This step isolates potential key phrases within the narrative text. Contrary to the traditional use of TF-IDF (Term Frequency-Inverse Document Frequency) metrics, RAKE employs a keyword score-matrix based on Word Frequency, Word Degree, and the Degree to Frequency Ratio. In the final phase, RAKE identifies the highest-scoring phrases among the phrase candidates. These phrases, representing the document's most significant themes or topics, are then used as the basis for student grouping, capturing the core interests manifest in the narratives.
International Research Journal of Modernization in Engineering Technology and Science, 2023
Fake job postings have become prevalent in the online job market, posing significant challenges t... more Fake job postings have become prevalent in the online job market, posing significant challenges to job seekers and employers. Despite the growing need to address this problem, there is limited research that leverages deep learning techniques for the detection of fraudulent job advertisements. This study aims to fill the gap by employing a Bidirectional Long Short-Term Memory (Bi-LSTM) model to identify fake job advertisements. Our approach considers both numeric and text features, effectively capturing the underlying patterns and relationships within the data. The proposed model demonstrates a superior performance, achieving a 0.91 ROC AUC score and a 98.71% accuracy rate, indicating its potential for practical applications in the online job market. The findings of this research contribute to the development of robust, automated tools that can help combat the proliferation of fake job postings and improve the overall integrity of the job search process. Moreover, we discuss challenges, future research directions, and ethical considerations related to our approach, aiming to inspire further exploration and development of practical solutions to combat online job fraud.
International Journal of Engineering Research & Technology (IJERT), 2022
Cardiac disease, which includes a variety of diseases that affect the heart, is a leading cause o... more Cardiac disease, which includes a variety of diseases that affect the heart, is a leading cause of death worldwide. One of every four deaths in the United States is due to heart disease. This means that approximately 610,000 people die from the disease each year. Heart disease is much easier to treat if detected early. Machine learning can play a crucial role in early detection and save lives. This research aims to develop an artificial intelligence-based system that identifies patients who are more likely to develop heart disease based on their medical history. The heart disease dataset from the UCI Machine Learning Repository was used for training and validation. Traditional classification techniques such as logistic regression, random forest, gradient boosting, and extreme gradient boosting were used as base models, and the results were compared with the Tabnet model. Tabnet is a new robust, interpretable, deeplearning architecture for tabular data. TabNet uses sequential attention to choose which features to conclude from at each decision step, focusing its learning ability on the most salient features, allowing for interpretability and more efficient learning. Promising results were obtained and validated using ROC curves, accuracy, precision, sensitivity, specificity, and confusion matrices. The Tabnet deep learning model outperformed the others, achieving 94% accuracy, ROC score of 0.94, and specificity and sensitivity greater than 0.93.
Journal of Intelligent Learning Systems and Applications, 2022
In this era of pandemic, the future of healthcare industry has never been more exciting. Artifici... more In this era of pandemic, the future of healthcare industry has never been more exciting. Artificial intelligence and machine learning (AI & ML) present opportunities to develop solutions that cater for very specific needs within the industry. Deep learning in healthcare had become incredibly powerful for supporting clinics and in transforming patient care in general. Deep learning is increasingly being applied for the detection of clinically important features in the images beyond what can be perceived by the naked human eye. Chest X-ray images are one of the most common clinical method for diagnosing a number of diseases such as pneumonia, lung cancer and many other abnormalities like lesions and fractures. Proper diagnosis of a disease from X-ray images is often challenging task for even expert radiologists and there is a growing need for computerized support systems due to the large amount of information encoded in X-Ray images. The goal of this paper is to develop a lightweight solution to detect 14 different chest conditions from an X ray image. Given an X-ray image as input, our classifier outputs a label vector indicating which of 14 disease classes does the image fall into. Along with the image features, we are also going to use non-image features available in the data such as X-ray view type, age, gender etc. The original study conducted Stanford ML Group is our base line. Original study focuses on predicting 5 diseases. Our aim is to improve upon previous work, expand prediction to 14 diseases and provide insight for future chest radiography research.
Uploads
Papers by Aravind Sasidharan Pillai