Browse Items (14428 total)
Sort by:
-
Machine Learning-Based Classification of Autism Spectrum Disorder across Age Groups
Autism Spectrum Disorder (ASD) is a complex neurodevelopmental condition that has gained significant attention in recent years due to its increasing prevalence and profound impact on individuals, families, and society as a whole. In this study, we explore the use of different machine learning classifiers for the accurate detection of ASD in children, adolescents, and adults. Furthermore, we conduct feature reduction to identify key features contributing to ASD classification within each age group using Cuckoo Search Algorithm. Logistic Regression has the highest accuracy compared to the other two models. 2024 by the authors. -
Machine Learning-Based Credit Scoring for Personalized and Inclusive Lending in Consumer-Centric Financial Systems
Traditional approaches to credit-scoring are largely based on rule systems that can be excessively fixed and limited to the ability to reflect individual financial behavior. The article analyzes the effectiveness of machine-learning (ML)- based credit ranking with the hypothesis that they can improve predictive capability and fairness of consumer credit lending. The performance of these algorithms, including supervised methods of learning, e.g., logistic regression, random forests as well as the deep learning, is contrasted to the conventional credit models. Model transparency is provided by SHAP values and other methods explainable by AI. Findings show that practice based on the use of ML outperform traditional methods in risk assessment, especially, through the inclusion of supplementary forms of data in traditional databases based on transaction behavior, virtual footprints, and psychometric signals. Furthermore, ethical standards and moral confidence in ML informed credit decision-making will require regulation-proof and explanatory modelling. Through the research, it is recommended to implement policy measures intended to cause financial institutions, fintech companies, and regulating bodies to implement ML-based credit-scoring technologies, with fairness and predictive effectiveness being reciprocal drivers of financial access and consumer-friendly lending practices. 2025 IEEE. -
Machine Learning-based Currency Information Retrieval for Aiding the Visually Impaired People
Paper currency is one of the most in-demand and long-established payment modes across the globe. People suffering from visual disabilities often face difficulties while handling paper currencies. Over the years, assisting technology has been rekindling itself to serve the aged and disabled person more aptly. Image processing methods and other sophisticated technologies, like Artificial Intelligence, Deep Learning, etc., can be employed to identify banknotes and fetch other valuable pieces of information from them. This paper proposes a framework that focuses on an integrated approach to retrieving data from the paper currency's uploaded image. The current version of the framework focuses on identifying the authenticity of the paper currency and classifying it according to its value. This work is an initiative to help visually impaired people to use paper currencies without assistance from other individuals and support them in living independently. 2021 IEEE. -
Machine Learning-Based Driver Assistance System Ensuring Road Safety for Smart Cities
Technologies around smart city and green computing are gaining more and more interest from diversified workforce areas. The transportation system is one of them. The transportation vehicles are operating day and night to provide proper support for the need. This is really tiring for the transportation workers, especially the drivers who are driving the vehicle. A slight negligence of a driver may cause a huge loss. The increasing number of road accidents is therefore a big concern. Research works are going on to comfort the drivers and increase the security features of vehicle to avoid accidents. In this chapter, a model is proposed, which can efficiently detect drivers drowsiness. The discussion mainly focuses on building the learning model. A modified convolution neural network is built to solve the purpose. The model is trained with a dataset of 7000 images of open and closed eyes. For testing purpose, some real-time experiments are done by some volunteer drivers in different conditions, like gender, day, and night. The model is really good for daytime and if the driver is not wearing any glass. But with a glass in the eyes and in night condition, the system needs improvements. 2025 selection and editorial matter, Yousef Farhaoui, Bharat Bhushan, Nidhi Sindhwani, Rohit Anand, Agbotiname Lucky Imoize and Anshul Verma; individual chapters, the contributors. -
Machine Learning-Based Imputation Techniques Analysis and Study
Missing values are a significant problem in data analysis and machine learning applications. This study looks at the efficacy of machine learning (ML) - based imputation strategies for dealing with missing data. K-nearest Neighbours (KNN), Random Forest, Support Vector Machines (SVM), and Median/Mean Imputation were among the techniques explored. To address the issue of missing data, the study employs k-nearest neighbors, Random Forests, and SVM algorithms. The dataset's imbalance is considered, and the mean F1 score is employed as an evaluation criterion, using cross-validation to ensure consistent results. The study aims to identify the most effective imputation strategy within ML models, offering crucial insights about their adaptability across various scenarios. The study aims to determine the best plan for data preprocessing in machine learning by comparing approaches. Finally, the findings help to improve our knowledge and application of imputation techniques in real-world data analysis and machine learning. 2024 IEEE. -
Machine Learning-Based Intrusion Detection Systems for 5G and beyond Networks
NextGen networks (5 G and beyond) have diversified their infrastructure. Traditional Intrusion Detection Systems (IDS) cannot effectively address the continuously evolving landscape of threats, which is why machine learning-based IDS has emerged as a crucial solution. This overview presents the trends in the application of machine learning techniques (deep learning and ensemble methods) for machine learning-based intrusion detection in 5 G and beyond networks. The important issues tackled encompass real-time anomaly detection, large-scale data processing, adaptive learning against unknown attacks, and detection outcomes. Specifically, we emphasize the promising combination of federated learning, reinforcement learning, and graph-based methods for deployment in distributed, resource-constrained network environments. We present a comprehensive overview of performance metrics such as accuracy, false positive rate, computational overhead, and scalability for each approach, highlighting the crucial trade-offs necessary for successful deployment in dynamic 5G scenarios. Furthermore, we prioritize privacy-preserving methods and secure model sharing. This abstract could further highlight that machine learning-based schemes for intrusion detection systems are important additions toward providing strong defences for cyberspace in 5 G and beyond. 2025 IEEE. -
Machine Learning-Based Maternal and Child Mortality Rate Prediction Using Random Forest Algorithm
This research uses a variety of data sources such as maternal age, health records of the mother and/or child, socioeconomic status, medical history, or prenatal care, and details of health indicators to determine the factors most decisive in increasing mortality risks. This entails data acquisition, data cleaning, data transformation and selection, and model building with an example of algorithms such as logistic regression and random forest. The trained models are checked for accuracy and their resilience level is checked using methods like SHapley Additive exPlanations and Local Interpretable Model agnostic Explanations for interpretation. The model is presented in an easy interface that doctors and health practitioners could use to make early and relevant decisions. It keeps updating the performance of established models and is a crucial way of maintaining data security for compliance with the set regulations. The rationale for this project is to offer practical recommendations for healthcare technicians so that more lives of mothers and children could be saved and maternal/child mortality decreased. Random Forest, in particular, has demonstrated superiority due to its ensemble approach, which mixes many decision trees to improve forecast accuracy and robustness. This technique can handle huge datasets with increased dimensionality and effectively lowers the overfitting risk. Additionally, Random Forest improves generalization by averaging the outputs of numerous trees, making it more tolerant to data noise and fluctuation. What makes it superior to single decision tree models is that it can handle both numerical and categorical data and handle missing values without a substantial loss of accuracy. 2025 selection and editorial matter, Babita Singla, Kumar Shalender, Nripendra Singh, and Sandhir Sharma; individual chapters, the contributors. -
Machine Learning-Driven Energy Management for Electric Vehicles in Renewable Microgrids
The surge in demand for sustainable transportation has accelerated the adoption of electric vehicles (EVs). Despite their benefits, EVs face challenges such as limited driving range and frequent recharging needs. Addressing these issues, innovative energy optimization techniques have emerged, prominently featuring machine learning-driven solutions. This paper reviews work in the areas of Smart EV energy optimization systems that leverage machine learning to analyse historical driving data. By understanding driving patterns, road conditions, weather, and traffic, these systems can predict and optimize EV energy consumption, thereby minimizing waste and extending driving range. Concurrently, renewable microgrids present a promising avenue for bolstering power system security, reliability, and operation. Incorporating diverse renewable sources, these microgrids play a pivotal role in curbing greenhouse gas emissions and enhancing efficiency. The review also delves into machine learning-based energy management in renewable microgrids with a focus on reconfigurable structures. Advanced techniques, such as support vector machines, are employed to model and estimate the charging demand of hybrid electric vehicles (HEVs). Through strategic charging scenarios and innovative optimization methods, these approaches demonstrate significant improvements in microgrid operation costs and charging demand prediction accuracy. The Authors, published by EDP Sciences, 2024. -
Machine Learning-Enabled NIR Spectroscopy. Part 3: Hyperparameter by Design (HyD) Based ANN-MLP Optimization, Model Generalizability, and Model Transferability
Data variations, library changes, and poorly tuned hyperparameters can cause failures in data-driven modelling. In such scenarios, model drift, a gradual shift in model performance, can lead to inaccurate predictions. Monitoring and mitigating drift are vital to maintain model effectiveness. USFDA and ICH regulate pharmaceutical variation with scientific risk-based approaches. In this study, the hyperparameter optimization for the Artificial Neural Network Multilayer Perceptron (ANN-MLP) was investigated using open-source data. The design of experiments (DoE) approach in combination with target drift prediction and statistical process control (SPC) was employed to achieve this objective. First, pre-screening and optimization DoEs were conducted on lab-scale data, serving as internal validation data, to identify the design space and control space. The regression performance metrics were carefully monitored to ensure the right set of hyperparameters was selected, optimizing the modelling time and storage requirements. Before extending the analysis to external validation data, a drift analysis on the target variable was performed. This aimed to determine if the external data fell within the studied range or required retraining of the model. Although a drift was observed, the external data remained well within the range of the internal validation data. Subsequently, trend analysis and process monitoring for the mean absolute error of the active content were conducted. The combined use of DoE, drift analysis, and SPC enabled trend analysis, ensuring that both current and external validation data met acceptance criteria. Out-of-specification and process control limits were determined, providing valuable insights into the models performance and overall reliability. This comprehensive approach allowed for robust hyperparameter optimization and effective management of model lifecycle, crucial in achieving accurate and dependable predictions in various real-world applications. Graphical Abstract: [Figure not available: see fulltext.]. 2023, The Author(s). -
Machine learning-enhanced heat and mass transfer study of elliptic motion in piezoelectric thermoelastic plates using Green-Naghdi III and three-phase-lag theories
Rayleigh-type surface waves in piezoelectric (PE) solids are pivotal for acoustic sensors, microelectromechanical systems (MEMS), and non-destructive evaluation. However, classical thermoelastic models fail under high heat flux due to the assumption of infinite thermal signal speeds, which limits their accuracy in coupled thermo-mechanical systems. To capture finite-speed and memory-dependent thermal effects, the Rayleigh wave propagation in a transversely isotropic (TI) PE half-space using generalized theories (such as Green-Naghdi type III (GN-III) and three-phase-lag (TPL)) is studied in this paper. The analytical formulation under varied electrical and thermal boundary conditions has been obtained. Secular equations are derived to characterize phase velocity, attenuation, and specific energy loss. A regression-based machine learning (ML) surrogate model is trained by using an analytical dataset to provide rapid predictions of wave parameters. Additionally, a confusion matrix classifier is applied to identify boundary conditions from simulated wave response features. The results demonstrated that the phase velocity increases with inclination angle and stabilizes with wave number, whereas attenuation and specific loss vary strongly by boundary condition (e.g., minimal in shorted-isothermal cases). The ML surrogate successfully reconstructed analytical predictions with minimal residual error, and the confusion matrix demonstrates accurate classification performance and validates the diagnostic potential of the framework. The novelty of this paper lies in integrating dual thermoelastic theories with machine learning, merging mechanics, heat transfer, and intelligent computing. These findings enable enhanced SAW sensor designs for precise gas/chemical detection, low-loss NDE tools for aerospace composite defect identification, and real-time diagnostics in biomedical ultrasonics for clearer imaging and efficient energy harvesting. 2026 Elsevier Ltd -
Machine Learning's Transformative Role in Human Activity Recognition Analysis
Human action recognition (HAR) is a burgeoning field of computer vision that seeks to automatically understand and classify the intricate movements performed by humans. From the graceful leaps of a ballerina to the decisive strides of a surgeon, HAR aims to decipher the language of motion, unlocking a plethora of potential applications. This abstract delves into the core of HAR, highlighting its key challenges and promising avenues for advancement. We begin by outlining the various modalities used for action recognition, such as RGB videos, depth sensors, and skeletal data, each offering unique perspectives on the human form. Next, we delve into the diverse set of algorithms employed for HAR, ranging from traditional machine learning techniques to the burgeoning realm of deep learning. We explore the strengths and limitations of each approach, emphasizing the crucial role of feature extraction and model selection in achieving accurate recognition. Challenges in Human Action Recognition (HAR), such as intra-class variations, inter-class similarities, and environmental factors. Ongoing efforts include robust feature development and contextual integration. The paper envisions HAR's future impact on healthcare, robotics, video surveillance, and augmented reality, presenting an invitation to explore the transformative world of human action recognition and its potential to enhance our interaction with technology. 2024 IEEE. -
Machine learningbased approaches for enhancing human resource management using automated employee performance prediction systems
Purpose: This study focuses on enhancing the accuracy and efficiency of employee performance prediction to enhance decision making and improve organisational productivity. By introducing advance machine learning (ML) techniques, this study aims to create a more reliable and data-driven approach to evaluate employee performance. Design/methodology/approach: In this study, nine machine learning (ML) models were used for forecasting employee performance: Random Forest, AdaBoost, CatBoost, LGB Classifier, SVM, KNN, XGBoost, Decision Tree and one Hybrid model (SVM + XGBoost). Each ML model is trained on an HR data set covering various features such as employee demographics, job-related factors and past performance records, ensuring reliable performance predictions. Feature scaling techniques, namely, min-max scaling, Standard Scaler and PCA, have been used to enhance the effectiveness of employee performance prediction. The models are trained to classify data, predicting whether an employees performance meets expectations or needs improvement. Findings: All proposed models used in the study can correctly categorize data with an average accuracy of 94%. Notably, the Random Forest model demonstrates the highest accuracy across all three scaling techniques, achieving optimise accuracy, respectively. The results presented have significant implications for HR procedures, providing businesses with the opportunity to make data-driven decisions, improve personnel management and foster a more effective and productive workforce. Research limitations/implications: The scope of the used data set limits the study, despite our models delivering high accuracy. Further research could extend to different data sets or more diverse organisational settings to validate the models effectiveness across various contexts. Practical implications: The proposed ML models in the study provide essential tools for HR departments, enabling them to make more informed data driven decisions with regard to employee performance. This approach can enhance personnel management, improve workforce productivity and fostering a more effective organisational environment. Social implications: Although AI models have shown promising outcomes, it is crucial to recognise the constraints and difficulties involved in their use. To ensure the fair and responsible use of AI in employee performance prediction, ethical considerations, privacy problems and any biases in the data should be properly addressed. Future work will be required to improve and broaden the capabilities of AI models in predicting employee performance. Originality/value: This study introduces an exclusive combination of ML models for accurately predicting employee performance. By employing these advanced techniques, the study offers novel insight into how organisations might transition from a conventional evaluation method to a more advanced and objective, data-backed approach. 2024, Emerald Publishing Limited. -
Machine learningbased approaches for enhancing human resource management using automated employee performance prediction systems
Purpose This study focuses on enhancing the accuracy and efficiency of employee performance prediction to enhance decision making and improve organisational productivity. By introducing advance machine learning (ML) techniques, this study aims to create a more reliable and data-driven approach to evaluate employee performance. Design/methodology/approach In this study, nine machine learning (ML) models were used for forecasting employee performance: Random Forest, AdaBoost, CatBoost, LGB Classifier, SVM, KNN, XGBoost, Decision Tree and one Hybrid model (SVM + XGBoost). Each ML model is trained on an HR data set covering various features such as employee demographics, job-related factors and past performance records, ensuring reliable performance predictions. Feature scaling techniques, namely, min-max scaling, Standard Scaler and PCA, have been used to enhance the effectiveness of employee performance prediction. The models are trained to classify data, predicting whether an employees performance meets expectations or needs improvement. Findings All proposed models used in the study can correctly categorize data with an average accuracy of 94%. Notably, the Random Forest model demonstrates the highest accuracy across all three scaling techniques, achieving optimise accuracy, respectively. The results presented have significant implications for HR procedures, providing businesses with the opportunity to make data-driven decisions, improve personnel management and foster a more effective and productive workforce. Research limitations/implications The scope of the used data set limits the study, despite our models delivering high accuracy. Further research could extend to different data sets or more diverse organisational settings to validate the models effectiveness across various contexts. Practical implications The proposed ML models in the study provide essential tools for HR departments, enabling them to make more informed data driven decisions with regard to employee performance. This approach can enhance personnel management, improve workforce productivity and fostering a more effective organisational environment. Social implications Although AI models have shown promising outcomes, it is crucial to recognise the constraints and difficulties involved in their use. To ensure the fair and responsible use of AI in employee performance prediction, ethical considerations, privacy problems and any biases in the data should be properly addressed. Future work will be required to improve and broaden the capabilities of AI models in predicting employee performance. Originality/value This study introduces an exclusive combination of ML models for accurately predicting employee performance. By employing these advanced techniques, the study offers novel insight into how organisations might transition from a conventional evaluation method to a more advanced and objective, data-backed approach. 2024 Emerald Publishing Limited -
Machine Learningcloud-Based Approach to Identify and Classify Disease
The term "Internet of Things"(IoT) describes the process of creating and modeling web-related physical objects across computing systems. IoT-based healthcare applications have offered multiple real-time products and benefits in recent years. For millions of people, these programmers provide hospitalization can get regular medical records and healthy lives. The introduction of IoT devices in the health sector has several technological developments. This study uses the IoT to construct a disease diagnostic system. Wearable sensors in this system initially monitor the patient's sympathy impulses. The impulses are then sent by a network environment to a server. In addition, a new hybrid approach to evaluation decision-making was presented as part of this research. This technique starts with the development of a set of features of the patient's pulses. Based on a learning approach qualifications are neglected. A fuzzy neural model was used as a diagnostic tool. A specific diagnosis of a particular ailment, such as the diagnosis of a patient's normal and abnormal pulse or the assessment of insulin issues, would be modeled to assess this technology. 2022 IEEE. -
Machine LearningEnabled NIR Spectroscopy. Part 2: Workflow for Selecting a Subset of Samples from Publicly Accessible Data
Abstract: An increasingly large dataset of pharmaceuticsdisciplines is frequently challenging to comprehend. Since machine learning needs high-quality data sets, the open-source dataset can be a place to start. This work presents a systematic method to choose representative subsamples from the existing research, along with an extensive set of quality measures and a visualization strategy. The preceding article (Muthudoss et al. in AAPS PharmSciTech 23, 2022) describes a workflow for leveraging near infrared (NIR) spectroscopy to obtain reliable and robustdata on pharmaceutical samples. This study describes the systematic and structured procedure for selecting subsamples from the historical data. We offer a wide range of in-depth quality measures, diagnostic tools, and visualization techniques. A real-world, well-researched NIR dataset was employed to demonstrate this approach. This open-source tablet dataset (http://www.models.life.ku.dk/Tablets) consists of different doses in milligrams, different shapes, and sizes of dosage forms, slots in tablets, three different manufacturing scales (lab, pilot, production), coating differences (coated vs uncoated), etc. This sample is appropriate; that is, the model was developed on one scale (in this research, the lab scale), and it can be great to investigate how well the top models are transferable when tested on new data like pilot-scale or production (full) scale. A literature review indicated that the PLS regression models outperform artificial neural network-multilayer perceptron (ANN-MLP). This work demonstrates the selection of appropriate hyperparameters and their impact on ANN-MLP model performance. The hyperparameter tuning approaches and performance with available references are discussed for the data under investigation. Model extension from lab-scale to pilot-scale/production scale is demonstrated. Highlights: We present a comprehensive quality metrics and visualization strategy in selecting subsamples from the existing studies A comprehensive assessment and workflow are demonstrated using historical real-world near-infrared (NIR) data sets Selection of appropriate hyperparameters and their impact on artificial neural network-multilayer perceptron (ANN-MLP) model performance The choice of hyperparameter tuning approaches and performance with available references are discussed for the data under investigation Model extension from lab-scale to pilot-scale successfully demonstrated Graphical Abstract: [Figure not available: see fulltext.]. 2023, The Author(s). -
Machine Transliteration of Handwritten MODI Script to Devanagari using Deep Neural Networks
The transliteration process involves transcribing words from the source language into the target language that uses a different script. Language and scriptural hurdles can be overcome via transliteration systems. There is a demand for automated transliteration systems due to the existence of several languages and the growing number of multilingual speakers. This study focuses on the Machine Transliteration of handwritten MODI script to Devanagari. MODI script was the official script for Marathi till 1950. Although Devanagari has, since then, taken over as the Marathi languages official script, the MODI script has historical significance as large volumes of its manuscripts are preserved in libraries across different parts of India. However, MODI into Devanagari transliteration is a difficult task because MODI script documents are complex in nature and there is no standard dataset available for the experiment. Machine Transliteration can be approached either as a Natural Language Processing task or as a pattern recognition task. In this research work, the transliteration task is carried out using the pattern recognition technique. The transliteration of MODI script to Devanagari is implemented using Convolutional Recurrent Neural Network (CRNN) based Calamari OCR, which is open-source software. An accuracy of 88.14% is achieved in character level matching of each word in the MODI to Devanagari transliteration process. When considering the entire word matching, the accuracy achieved is 61%. Machine Transliteration of MODI script documents results in the retrieval of large repositories of knowledge from ancient MODI manuscripts. (2024), (Research Institute of Intelligent Computer Systems). All rights reserved. -
Machine-Learning Based Sleep Pattern Analysis Using Linear Regression Algorithm
This article is investigating the connection between sleep patterns and concentration spans among university students while exploring the potential influence of MyersBriggs Type Indicator (MBTI) personality types on these aspects. The primary objective is to understand how sleep duration affects students ability to maintain focus and how their personality traits might interact with this relationship. Data was collected from university students aged 1619 using a multiple-choice form. The key variables analyzed were age, MBTI personality types, sleep duration, concentration span, and effective study ranking. Pearson's correlation was employed to examine these relationships. Additionally, a linear regression model was developed to predict concentration span based on sleep hours. The findings revealed a strong positive correlation 0.758 between sleep duration and concentration span, suggesting that increased sleep is associated with longer concentration spans. A moderate positive relationship 0.249 was also observed between concentration span and effective study ranking. However, the analysis showed a negligible relationship ? 0.008 between MBTI personality types and concentration span, indicating that within the context of this study, personality type does not significantly influence concentration span. This research emphasizes the critical role of sleep in academic settings and challenges the assumption that personality types significantly impact concentration span and sleep patterns. The linear regression model developed provides a predictive tool for understanding the impact of sleep on concentration, underscoring the importance of adequate sleep for academic success. This research is contributing to the broader understanding of factors influencing student performance and offers practical insights for optimizing study habits and educational strategies. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Machining Characteristics Evaluation of Al7075TiB2 In Situ Composite Using Abrasive Water Jet Machining with Varied Test Parameters
The study delves into the abrasive water jet (AWJ) cutting of an Al7075TiB2 metal matrix composite that was synthesized in situ. The primary goal is to investigate how variations in three key process parameters, namely, stand-off distance (SOD) ranging from 0.5 to 2.5mm, abrasive flow rate (100 to 300gmin), and traverse speed (100 to 500mmmin), affect three critical performance metrics: volumetric material removal rate (VMRR), dimensional accuracy, and surface roughness (SR). The study's findings were represented graphically, highlighting the relationships between these responses and the aforementioned process parameters. Scanning electron microscopy (SEM) was also used to examine the machined surfaces. It was discovered that increasing traverse speed resulted in significant increases in surface roughness, VMRR, and dimensional errors. An increase in the SOD, on the other hand, resulted in an increase in surface roughness, VMRR, and a decrease in dimensional accuracy. Furthermore, increasing the abrasive flow rate resulted in lower surface roughness and dimensional accuracy while achieving a higher VMRR. 2023, The Institution of Engineers (India). -
Machining Characteristics Evaluation of Al7075TiB2 In Situ Composite Using Abrasive Water Jet Machining with Varied Test Parameters
The study delves into the abrasive water jet (AWJ) cutting of an Al7075TiB2 metal matrix composite that was synthesized in situ. The primary goal is to investigate how variations in three key process parameters, namely, stand-off distance (SOD) ranging from 0.5 to 2.5 mm, abrasive flow rate (100 to 300 g min), and traverse speed (100 to 500 mm min), affect three critical performance metrics: volumetric material removal rate (VMRR), dimensional accuracy, and surface roughness (SR). The studys findings were represented graphically, highlighting the relationships between these responses and the aforementioned process parameters. Scanning electron microscopy (SEM) was also used to examine the machined surfaces. It was discovered that increasing traverse speed resulted in significant increases in surface roughness, VMRR, and dimensional errors. An increase in the SOD, on the other hand, resulted in an increase in surface roughness, VMRR, and a decrease in dimensional accuracy. Furthermore, increasing the abrasive flow rate resulted in lower surface roughness and dimensional accuracy while achieving a higher VMRR. The Institution of Engineers (India) 2023. -
MADeGen: Multi-Agent based Deep Reinforcement Learning for Sequential Keyphrase Generation
Keyphrase generation is an essential tool in the field of natural language processing for information retrieval, document summarization, and text recommendation applications, extracting succinct and representative phrases from the text document. Traditional keyphrase extraction methods applied the supervised or unsupervised learning fail to capture the sequential keyphrase generation in a dynamic environment. The keyphrase generation approaches lack focus on explicitly discriminating the present and absent keyphrases, leading to the inadequate generation of semantically rich absent keyphrases. Hence, this work utilizes the potential benefits of reinforcement learning with the design of a distinguished reward function for present and absent keyphrases for sequential decision-making in the keyphrase generation. Thus, this work presents a novel keyphrase generation system, MADeGen, utilizing Multi- Agent Deep Reinforcement Learning (MADRL). In particular, a multi-agent reinforcement system collaboratively enables the generation of representative and coherent keyphrases by the evaluation metric-aware cooperative reward function analysis and adaptively training the agents. The proposed MADeGen incorporates two major phases, such as multi-agent modelling and actor critic-based policy optimization towards accurate keyphrase generation. In the first phase, the proposed approach designs two learning agents, including the extraction agent and generation agent, with the incorporation of a pre-trained language model. In the multi-agent system, the generation agent is the finetuned version of the extraction agent with the integration of the Wikipedia source. Secondly, the evaluation-aware adaptive reward function is designed to evaluate each agent's generated keyphrases with reference to ground-truth keyphrases. In subsequence, the cooperative reward analysis triggers the actor critic-based policy optimization for the generation agent in the multi-agent system to precisely generate the semantically relevant keyphrases with the assistance of an external web source. Experimental results on several benchmark datasets, such as Inspec, PubMed, and wiki20, illustrate the effectiveness of the proposed MADeGen compared to the existing keyphrase extraction models, yielding state-of-the-art performance in keyphrase extraction tasks. The proposed MADeGen proves its higher performance in the present as well as absent keyphrase extraction as 0.367 and 0.438 F1-score, respectively, while testing on the Inspec dataset. (2024), (Intelligent Network and Systems Society). All Rights Reserved.