International Journal on Recent and Innovation Trends in Computing and Communication

A new Dynamic Routing Approach for Software Defined Network

Saroj Singh, Kamlesh Sharma — Fri, 10 Jan 2025 00:00:00 +0000

Introduces a new dynamic routing approach tailored for Software Defined Network (SDN) that takes advantages of the programmability and centralized control inherent SDN architectures. Traditional routing protocols often struggles often to adapt to dynamic network conditions, leading to suboptimal performance and resource utilization. In contrast the objective of the paper is to proposed approach uses real time network information collected by the SDN controller to dynamic adjust routing decisions and dynamic routing algorithms for software define networks in wide area network (SDN-WAN), provide a new approach; By employing a combination of machine learning algorithm and network speed back mechanism. Using the approach optimizes routing paths based on factors such as link utilization and quality of service requirements. The shortest feasible path (SOFP) is an adaptation of the shortest feasible path algorithm that uses a statistical technique from the OpenFlow interface. The goal of the SFOP algorithm is to efficiently use SDN-WAN resources by determining the best route from source to destination. Overall, the dynamic routing approach provides a promising solution to efficiently manage network traffic in SDN. Paving the way for more adaptative and responsive networking infrastructure.

Multi-Cloud Observability: Tools and Techniques for Monitoring and Troubleshooting Complex Hybrid Cloud Environments

Arun Pandiyan Perumal, Viralkumar Ahire — Fri, 10 Jan 2025 00:00:00 +0000

This article focuses on detection tools and methods for hybrid cloud that are used to deal with complexity levels within multi-cloud infrastructures. It breaks down some of the best open-source and commercial observability solutions, like Prometheus, Grafana, Jaeger, Datadog, New Relic, Dynatrace, and Splunk, describing the offered functions, their advantages, and disadvantages. Some of the problems highlighted in the research include multi-cloud visibility and data consistency, integration difficulties, and inherent scalability. The real-life examples of TD Bank and Blinkit show how organizations can use and realize the values of observability solutions for better service dependability, quick reaction to incidents, and customer satisfaction. The paper then analyses some of the new trends, like the use of artificial intelligence in monitoring and enabling automated repairs and improvements to the network while at the same time trying to improve operational efficiency and looking at operation costs. Core problem-solving approaches for multi-cloud cases are articulated, which include diagnostics of the root cause, proper handling of the incident handling process, and the use of intelligent automation for problem-solving. Thus, the results highlight the need to implement extensive observability strategies for the effective management of distributed cloud systems. Future advancements are expected with cloud technologies; hence, organizations need to keep abreast of the latest concerning observability tools and approaches to ensure their multi-cloud environments remain high on performance and reliability.

Spectroscopic and Chromaticity Examination of Disappearing Ink

Parvesh Sharma, Kapil Kumar, Himanshu A. Pandya — Fri, 10 Jan 2025 00:00:00 +0000

The increasing prevalence of disappearing ink pens in forgery cases has posed significant challenges to forensic document examiners. This study investigates disappearing ink's spectroscopic and chromaticity characteristics to establish a reliable method for determining the relative age of writing. Using a Video Spectral Comparator (VSC-6000/hs), the absorbance, reflectance, and fluorescence spectra of disappearing ink were analyzed over time. Chromaticity values were evaluated within the CIE L*a*b* color space to quantify changes in the ink's optical properties as it faded. Results demonstrate that the L* value, representing luminance, showed a strong correlation with time in absorbance and reflectance spectra (R² = 0.82 and 0.81, respectively). Statistical analysis confirmed the linearity of L* values, making it a significant indicator for estimating the time elapsed since writing. The study concludes that absorbance and reflectance measurements are more effective than fluorescence for analyzing disappearing ink. This research provides forensic experts with a sensitive, non-destructive, and reproducible method to analyze disappearing inks and estimate the relative time of writing, offering a valuable tool for addressing forgery-related challenges.

Cardiovascular Disorder Detection in Diabetes Mellitus Patients: An Integrated VGG and Bi-LSTM Model Optimized Using the ABC Algorithm

Bhagyalaxmi, Muktevi Srivenkatesh — Fri, 10 Jan 2025 00:00:00 +0000

There is a major public health concern at the intersection of Diabetes Mellitus (DM) and Cardiovascular Diseases (CVDs). Patients with a diabetes diagnosis are more likely to experience a variety of cardiovascular problems. Better patient outcomes and lower healthcare costs can result from early diagnosis of these problems. This study presents a fresh computational model to tackle this problem. This research presents an integrated method that optimizes the VGG and Bidirectional Long Short Tem Memory (Bi LSTM) models together with the help of the Artificial Bee Colony (ABC) algorithm, which is based on the swarm intelligence of artificial bees. Cardiac images are processed using the VGG network, which has been shown to be highly effective in image classification, while the Bi LSTM is optimized for processing time series data from medical sensors, such as heart rates and blood sugar levels. The selected characteristics are then used in the proposed VGG 16 model before being sent to Bi-LSTM for further processing and abnormality detection. The VGG consists of 16 layers, all of which are blocks of 2D Convolution and Max Pooling layers. The ABC method was created as a result of research into intelligent behavior and is now widely used in areas such as problem solving, categorization, and optimization. The ABC algorithm is used to the unified model, which results in improved adaptability, speed of convergence, and robustness. To better forecast cardiovascular diseases, this research presents an Integrated VGG16 model with Bi-LSTM model with ABC optimization (VGG-Bi-LSTM-ABC) to predict the cardiovascular disorders. When compared to the standard model, the proposed model's ability to detect disorder is much better. Preliminary results from a carefully selected dataset of DM patients show that the integrated model outperforms state-of-the-art approaches in key measures, further demonstrating the promise of Artificial Intelligence (AI)-driven advances in medical diagnosis.

Secure Data Transmission to Improve the Performance of Communication in Hybrid Systems

Prajakta A. Satarkar, Girish V. Chowdhary — Fri, 10 Jan 2025 00:00:00 +0000

Light Fidelity (Li-Fi) is a way of communication using LED’s with a high data rate and secures data transmission. But it has some drawbacks like data loss during shadowing and flickering of light, interference, etc. To deal with data loss we can use Hybrid Li-Fi and Wi-Fi Networks (HLWNets) by combining Li-Fi and Wireless Fidelity (Wi-Fi). Emerging HLWNets design and implementation face secure data transmission issues introduced due to Wi-Fi networks. We propose a novel comprehensive solution called Efficient Handover Protocol with Secure Data Transmission (EHPSDT).To assure the total security of data, we proposed security architecture based on Attributed-based Elliptic Curve Encryption (AECC) that ensures confidentiality and integrity. It also allows for fine-grained access control in HLWNets. Compared to other current methodologies, the proposed method minimizes overall processing overhead. The result of simulation revealed the performance of the proposed EHPSDT compared to underlying methods in terms of packet delivery ratio (PDR), average throughput, and communication overhead.

Innovativeness, Skill Development, Competitive Efficiency, Capacity of Hard Work and Entrepreneurial Intentions among the Students of Higher Learning Institutions - An Assessment

G. Yoganandham — Fri, 10 Jan 2025 00:00:00 +0000

This research explores the dimensions of innovativeness, skill development, competitive efficiency, capacity for hard work, and entrepreneurial intentions among students of higher learning institutions in Tamil Nadu. In the context of a rapidly evolving job market and economic landscape, these attributes are critical in shaping the future workforce. The research aims to identify the factors that influence these traits and their interrelationships, contributing to a holistic understanding of student readiness for entrepreneurial and professional careers. Information was collected through a view of students across various disciplines in higher learning institutions, including engineering, management, and arts and sciences. The research employed valued performing to assess the levels of innovativeness and entrepreneurial intentions among students, along with soft insights to understand the underlying motivations and barriers. The accumulation indicate that innovativeness is significantly associated with exposure to practical learning experiences and a supportive academic environment. Skill development was found to be influenced by access to advanced training and mentorship, while competitive efficiency correlated with participation in extracurricular activities and internships.

The capacity for hard work, though generally high, varied according to personal resilience and external support systems. Entrepreneurial intentions were notably higher among students who demonstrated a strong alignment between their skills and market needs, as well as those who had role models in entrepreneurship. However, challenges such as lack of financial resources, inadequate institutional support, and societal expectations were identified as major obstacles to entrepreneurial pursuits. The research suggests that educational institutions should incorporate entrepreneurship education, real-world problem-solving opportunities, and skill development support systems to equip students for competitive ventures and economic growth in Tamil Nadu.This research underscores the importance of a multidimensional approach to education that not only imparts knowledge but also cultivates the essential qualities needed for innovation and entrepreneurship in a dynamic global economy.

Internet of Things (IoT) Adoption in Higher Education Institutions: An Empirical Study in Saudi Arabia Universities

Jebreel Alamari — Fri, 31 Jan 2025 00:00:00 +0000

The Internet of Things (IoT) may offer many advantages to academic institutions, but its adoption, like other technologies, may also result in unanticipated risks and the necessity of significant organizational adjustments. This study examines the adoption of IoT by Saudi public and private universities. It targets the students and teachers to measure their intentions and actual behaviors to adopt IoT in academic research. An exhaustive literature review is necessary to create the research hypotheses and classify the anticipated benefits and risks of the Internet of Things (IoT). For the purpose of gaining an understanding of the relationships between university and technology, the study offers a theoretical framework by developing research hypotheses. The study used a quantitative research design by administering the survey questionnaires among the students and teachers of 7 public and private universities in Saudi Arabia. The study received 338 filled responses from the survey questionnaires. The findings showed that perceived usefulness and ease of use significantly and positively influence the intention to adopt IoT. Additionally, perceived ease of use significantly and positively influences perceived usefulness. Finally, the study found that the intention to adopt IoT significantly and positively influences actual user behavior to adopt IoT in academic research. The study recommends that the internet of things (IoT) may then provide universities with a multitude of benefits. It is necessary to make modifications to the organization, its procedures, and its systems to cultivate capabilities and make sure that IoT is compatible with the objectives of academic institutions

Literature survey on Feature Extraction methods using CBIR Visual Search

S. Pratap Singh, Ch. Bindu Madhuri, P. Satheesh — Fri, 31 Jan 2025 00:00:00 +0000

Efficient image retrieval relies on robust feature extraction methods capable of capturing the distinct characteristics of color, texture, and shape. This study investigates diverse techniques across these domains, emphasizing their impact on the ranking accuracy of retrieved images. In the color domain, methods such as Color Moments (CM), Color Moment Invariant Model (CMI), Dominant Color-Based Vector Quantization (DCVQ), MPEG-7 Dominant Color Descriptor, and integrated color-texture approaches are explored for their precision in identifying chromatic variations. Texture feature extraction techniques, including Discrete Wavelet Transform (DWT), Statistical Edge Detection (SED), Modified Scalable Descriptor (MSD), and Local Derivative Radial Patterns (LDRP), alongside Support Vector Machine (SVM) classifiers, are assessed for their ability to identify and rank images based on structural complexity. For shape features, advanced techniques such as boundary moments, complex coordinates, curvature scale space, intersection point mapping, and merging strategies are evaluated for their role in preserving spatial and geometric fidelity. By examining these methods in the context of top-ranked image retrieval, this work provides a comparative framework to guide the selection of optimal feature extraction techniques for high-performance image analysis systems.

Smart Conferencing Rooms: A Comprehensive Approach to AI-Driven Gesture and Virtual Interaction

Fri, 31 Jan 2025 00:00:00 +0000

The “Smart Conferencing Rooms” project is aimed at changing the interaction of users in educational and healthcare facilities through the use of artificial intelligence in hand gesture recognition and virtual communication technologies. By embedding Artificial Intelligent Hand Gesture Recognition and Virtual Communication Technologies to enhance the user interaction in education and health care these are the objectives of the “Smart Conferencing Rooms”. The goal is to create new AI/ML models for hand detection, fingertip tracking and air writing allowing for such a realization of interactions as drawing on the midair, typing, etc. , or solving mathematical problems. Further, it aims at improving more real-time interaction through the virtual chat system with better face and gesture recognition than Google Meet but with more functions. This cross-platform solution is asserted to enhance the qualities of communications and enhance collaboration and connectivity, which will be highly beneficial to these sectors.

This work is inspired from progress made in the field of object tracking which is one of the core issues of computer vision and it entails the ability to recognize and find objects like hand gestures in successive frames. It is suitable for those applications such as automatic surveillance and video recognition. By analyzing gestures the system can translate them into text which will especially be so beneficial to the deaf group of people because it helps them communicate effectively using gestures.

Evaluating the TOE Framework for Technology Adoption: A Systematic Review of Its Strengths and Limitations

Chandra Prakash — Tue, 18 Feb 2025 00:00:00 +0000

The adoption and decision-making of information technology (IT) remains the cornerstone of organizational innovation and market competitiveness. Various frameworks, such as the Technology Acceptance Model (TAM), the diffusion of innovations (DoI), and the Technology-Organization-Environment (TOE) framework, have been utilized to explain IT adoption decision-making. Among these, the TOE framework stands out for its holistic approach. The TOE framework has demonstrated adaptability across industries and technologies and has been used to examine technological capabilities, organizational readiness, and environmental influences on technology adoption. However, there remains a persistent debate about the TOE framework’s theoretical rigor and contextual applicability to address decision-making about technology adoption. This systematic review critically analyzes the strengths and limitations of the TOE framework while comparing and contrasting it with the DoI and TAM frameworks for technology adoption. This paper identified the gaps, such as the limited consideration of dynamic adoption processes and post-adoption outcomes in the TOE framework. This research synthesizes existing knowledge and critiques the current utility of the framework. It also offers a foundation for its evolution, addressing a significant scholarly need for critical evaluation and innovation in technology adoption studies.

Federated Learning a Collaborative Machine Learning Across Countries with Data Privacy

Narendra Lalkshmana Gowda — Mon, 03 Mar 2025 00:00:00 +0000

With growing importance of data in shaping policies, economic strategies, and healthcare systems, securing citizens data has become a critical issue for national governments. At the same time, the potential benefits of large-scale collaborative machine learning (ML) across countries are undeniable. Federated learning (FL) offers a unique solution to this dilemma by enabling the training of AI models across decentralized data sets without requiring data to be shared. This paper explores how different countries can use federated learning to contribute to collaborative machine learning while ensuring national data security. We examine the privacy-preserving mechanisms in FL, the technical challenges, and propose a framework for cross-country collaboration on a global scale..

Lora based Manhole Cover Status and Toxic Gas Monitoring with IoT Technologies

Meltem Eryılmaz, Aslıhan Bilge Şener, Yasin Cetinkaya, Mehmet Özdem — Mon, 03 Mar 2025 00:00:00 +0000

Currently smart city management is of critical importance in the urban infrastructure of growing countries. In this context, it is necessary to make the underground manholes and sewage systems smart and traceable. Open manhole covers pose security risks which may lead to various accidents and damages. Gas leaks, on the other hand, pose serious risks and endanger human health. In this study, a wireless device was developed to monitor the condition of manhole covers and track gas levels within them. This device, automatically monitors manhole gases, detects gas leaks, checks the status of the covers and through the Internet of Things (IoT) system it is connected to provides alarms to inform the authorities. The sensors used inside the device can detect the presence of toxic gases, thereby contributing to occupational health and environmental safety.

Optimized AES with GAN Model for Secure Medical Image Transmission

Vishnupriya V, Sarathambekai S — Mon, 24 Mar 2025 00:00:00 +0000

The rapid technological development and increased computational capabilities, cybersecurity risks are on the rise. This has led to a growing need for cutting-edge security algorithms, especially in fields like healthcare where medical images play a crucial role in diagnosing various conditions. As these images are frequently transmitted over the internet, safeguarding them from cyber threats is essential. The new framework for encryption is named PSO-AES-GAN(PSAGA). This paper introduces PSO based AES for encryption and generative GAN (Generative Adversarial Network) for key generation to strengthen the security of medical images. The model leverages an AES with PSO (Particle Swarm Optimization) encryption, SHA- 256 hash table, and GAN deep learning techniques. A SHA- 256 hash-table-based equation and AES with PSO enhance key entropy. Differential Huffman Compression (DHC) is utilized to compress encrypted images low-loss. The medical images have undergone testing using this model and assessed using performance metrics such as entropy, Encryption time, Decryption time, and Compared encryption algorithms such as chaotic maps, DES, AES, and Blowfish with similarity. Results show that the suggested model outperforms current methods.

AI Agents for Business Applications: A Review

Palguni G T, Thyagaraju G S — Mon, 24 Mar 2025 00:00:00 +0000

Artificial Intelligence (AI) agents have revolutionized business applications by automating processes, enhancing decision-making, and optimizing operational efficiency. This paper presents a comprehensive review of AI agents, categorizing their applications across domains such as customer relationship management (CRM), supply chain management, financial forecasting, and enterprise decision support systems. The evolution of AI agents from rule-based models to sophisticated multi-agent systems (MAS) and large language models (LLMs) has enabled businesses to leverage intelligent automation, real-time analytics, and predictive insights. AI-driven conversational agents have improved customer engagement, while AI-powered workflow automation has enhanced IT operations and security. Despite these advancements, challenges such as ethical considerations, security risks, interoperability, and long-term adaptability persist. This review synthesizes research contributions, identifying key strengths, limitations, and emerging research gaps in AI adoption for business. Future directions highlight the need for enhanced human-AI collaboration, standardization of AI agent interoperability, security-first AI architectures, and emotionally intelligent conversational systems. Addressing these challenges will ensure the responsible and effective deployment of AI agents, maximizing their transformative potential in business environments.

Securing Digital Identities System through Blockchain Networks

T Vairam, M Srijeimathy, Mukilan A — Wed, 21 May 2025 00:00:00 +0000

The increasing importance of blockchain technology as an improvement in efficiency, security, and transparency across different fields has notably made a difference in identity and access management systems. Traditional security don’t extend sufficient protection, mainly because threats have become sophisticated. Decentralisation and cryptographic mechanisms assure the integrity of data and eliminate single points of failure offered by a blockchain. This paper highlights the transformative role of blockchain technology in cybersecurity, especially with focus on smart contracts, secure authentication techniques, and decentralised identity management. The paper provides an insight into technological advancements, challenges, and trends with respect to bolstering security and trust in online transactions. It goes a step further to evaluate the viability of implementing blockchain-based identity management systems by the corporate world and governmental organisations and takes into account factors including scalability, regulatory compliance, and user adoption.This paper also proposes blockchain based vehicle identity system using Practical Byzantine Fault Tolerance (PBFT) and Directed Acyclic Graph (DAG ) Consensus

Smart Irrigation System using Raspberry Pi

Sneha S. Temgire, Y. S. Angal — Wed, 21 May 2025 00:00:00 +0000

This project focuses on automated and manual irrigation in addition with plant disease detection and growth monitoring using image processing on a Raspberry Pi 3B+. By leveraging TensorFlow Lite and OpenCV, the system can analyze plant health and trigger appropriate irrigation actions. The aim is to design accurate agriculture system by reducing water wastage and improving crop monitoring.

Heart Disease Prediction using Integrated Technology of XGBoost, Random Forest and Multi-Layer Perceptron

Wed, 21 May 2025 00:00:00 +0000

Cardiovascular disease remains a leading cause of death worldwide, requiring prompt and accurate diagnosis to minimize patient mortality rates. More recent developments in artificial intelligence (AI) applications have demonstrated how to enhance prognostic performance and interpretability in clinical diagnosis. This research paper analyzes the application of machine and Deep Learning models for heart disease prediction by voting with a selection of models in order to develop a strong classifier. A weighted ensemble voting approach is employed and leverage is made from XGBoost, Random Forest, and Multi-Layer Perceptron (MLP) model strengths. Further, explainability is offered by SHapley Additive exPlanations (SHAP) to facilitate model decisions, allowing feature importance and decision-making insight. The proposed methodology is supported by established performance metrics, retaining clinical relevance. Results imply that AI-based approaches can achieve elevated predictive accuracy and interpretable diagnoses, informing the creation of automated cardiovascular risk stratification.

Artificial Neural Networks and Optimization Technique: A theoretical study

Basir Ahamed Khan — Mon, 02 Jun 2025 00:00:00 +0000

Artificial Neural Networks (ANNs) have become a pivotal tool in modern artificial intelligence (AI), significantly impacting various fields such as image processing, natural language processing, and autonomous systems. The training process of ANNs requires find-ing optimal parameters (weights and biases) that minimize a loss function, which can be computationally intensive and challenging. To achieve better performance, it is crucial to employ efficient optimization techniques that guide the network toward optimal solutions effectively. This paper provides an overview of ANNs, including their structure, types, applications, advantages, challenges, and future directions. This review also provides optimization techniques that are used to enhance their performance during training.

Machine Learning-based Intrusion Detection System for Social Network Infrastructure

Govind Kumar Jha, Preetish Ranjan, Ritesh Ravi, Hardeo Kumar Thakur — Fri, 27 Jun 2025 00:00:00 +0000

The growing number of cyber-attacks demands a critical measure to prevent unauthorized data access. Thus, intrusion detection has become critical to deal with such attacks. This work attempts to identify malicious connections using a few key parameters. The system has been trained using data relating to normal and abnormal events through machine learning and data mining techniques. To detect intrusions, this study assessed five distinct machine learning models: Random Forest, Bagging, Boosting, Support Vector Machine, and K-Nearest Neighbor (KNN). Based on the number of features, iterations, and hyperparameters, the models were evaluated using experimental data collected in real time. With a detection rate of up to 98.7%, the Random Forest approach surpassed existing machine learning models for intrusion detection. The paper proposes a novel intrusion detection system (IDS) based on these findings that successfully identifies possible threats before they seriously compromise network security and stop cyberattacks.

Skin Disease Classification Using Multi-Model Optimization and Augmentation

Fri, 27 Jun 2025 00:00:00 +0000

Skin diseases affect millions globally, posing screening challenges due to complex lesion characteristics and limited access to medical expertise. Traditional screening methods are time consuming, often requiring extensive laboratory testing. Deep learning and machine learning techniques have gained significant traction in recent years, serving as powerful tools in tackling complex problems, particularly in areas requiring substantial prior knowledge, such as biomedicine. With the challenge of inadequate medical resources, these methods have found impactful applications in disease screening, emerging as a pivotal research focus on dermatology. This project aims to develop an automated skin disease screening system using machine learning and deep learning techniques. The system is designed to accurately identify skin diseases, enhance early detection, address existing challenges in screening and ensure accessibility and affordability for all. This provides a concise review of the classification of skin diseases, leveraging Convolutional Neural Networks (CNN) and K-Nearest Neighbors (KNN) to analyze skin lesion characteristics and evaluate imaging technologies. By exploring the strengths of CNNs due to its high performance in image classification and feature extraction. KNN providing evidence by identifying similar images, making it an explainable AI model. This study presents an Evidence based screening system a virtual dermatology platform leveraging cutting-edge artificial intelligence and deep learning techniques for efficient skin disease classification. Using pre-trained models like GoogleNet, EfficientNet, ResNet, DenseNet, MobileNet and achieving a classification accuracy of 97% through EfficientNet. significantly reducing screening time and cost. The proposed system optimizes preprocessing, transfer learning, model training and cross-validation, significantly improving accuracy. The results highlight AI's potential to revolutionize dermatological screening, reducing costs and improving early detection.

Self-Supervised Hierarchical Representation Learning for Multi-Dimension Context

Manali Jahagirdar, Mukta Takalikar — Fri, 27 Jun 2025 00:00:00 +0000

Self-supervised hierarchical representation learning offers an effective approach to capturing multi-dimensional context from unlabeled data. A key challenge in representation learning is integrating information from diverse aspects of the input, particularly when labeled data is limited. To address this, a novel strategy can be introduced that learns representations hierarchically, enabling the capture of context at varying levels of abstraction and across multiple dimensions. The process begins by modeling different contextual facets through component-specific representations, each capturing distinct semantic and structural attributes. A dynamic aggregation mechanism then combines these representations in a hierarchical manner, allowing information to propagate across levels of contextual abstraction. This enables the encoding of both fine-grained nuances and broader contextual dependencies. By leveraging self-supervised learning, the approach optimizes for inherent relationships within the multi-dimensional context, enabling the acquisition of robust representations from unlabeled data. This makes it particularly suitable for domains where labeled data is scarce or costly to obtain. Experimental results highlight the ability to learn rich, hierarchical representations that enhance performance on downstream tasks requiring deep contextual understanding. Key technical contributions include: (1) a context-aware masking strategy using Text Encoder for semantic recovery of masked fields, (2) a Hierarchical Model that fuses fine-grained tabular features with coarse-grained concepts and (3) a multi-stage training code base combining contrastive loss for cross-document alignment (RFP-bid pairs) and silhouette scores (from scikit-learn) to validate cluster coherence.

Retracted

Mon, 07 Jul 2025 00:00:00 +0000

Retracted

Performance evaluation of ovarian cancer detection using on machine learning approaches based on feature selection

D. Savitha, D. Rajakumari — Sat, 12 Jul 2025 00:00:00 +0000

Ovarian cancer is one of the most dangerous genecology cancers because it does not show any symptoms in the early stages and there aren’t any good tests to find it. Early detection is important for increasing patient survival, but traditional diagnostic methods often don't have the right level of sensitivity and specificity. This study looks into how machine learning (ML) and deep learning (DL) can be used to find ovarian cancer early and accurately. We looked at five models: Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), Artificial Neural Network (ANN), and Elman Recurrent Neural Network (ERNN). We did this before and after using the RF algorithm to select features. The results show that feature selection made all of the models work much better. The ERNN model performed the best overall, with accuracy going from 89.8% to 92.5% and AUC-ROC going from 0.94 to 0.96 after feature selection. In the same way, ANN and RF got 92.1% and 91.0% accuracy, respectively, with big improvements in precision, recall, and F1-score. These results show how important it is to optimize features to make models work better. They also confirm that intelligent ML-based systems could be used to reliably find ovarian cancer early.

Advances in Prompt Engineering and Retrieval-Augmented Generation for Scalable AI Systems

Mon, 14 Jul 2025 00:00:00 +0000

Immediacy recently become a hot topic of scalable AI system and technologies, due to the rapid development in?AI, especially in NLP. “Noising” Prompt Writing The goal of effective prompt design is to write an input prompt, or set of prompts, that?help encourage LLMs to produce the desired output given the context and in contrast to other output. Approaches like the automatic and flexible prompt generation, few-shot learning, transfer learning to specific domains without?needing to re-train the models below, etc., made the prompts become dominant as an interface in the big model era. Manifesting this aim for fast engineering, retrieval-augmented generation is an instantiation?of the transfer of outside data to instantaneously influence the generation. As opposed to static material from document stores or databases which refines the answer with the latest and most?correct information, classic LLMs condition the answer on massive pre-trained knowledge. The hybridisation of these analogue knowledge sets serves to exploit the strengths of the two, and so?this is a more effective than the previous method of taking each one in isolation. A more efficient and precise AI would be possible by integrating the?two successics so that we have a trustworthy Dialogue system, decision support, and etc. Fast querying strategy, adaptive algorithms, and modular design?for interacting just in time with low intensity calculation are the key technology innovations. However, there are still several challenges that need to be addressed to make prompt-based design more reliable, handle retrieval noise, trade off latency and quality and use?it responsibly to mitigate bias and disinformation. Decentralised retrieval for better privacy and scalability, and multimodal retrieval and generation that could self-optimise using reinforcement learning, are some of?the interesting directions to explore in the future. We also?illustrate the interplay between these two relatively new developments in AI system design: retrieval-augmented generation and blitz engineering. In it you will find benchmarking?performance, current trends, and best practices that elevate AI from static information to dynamic knowledge through responsive, context-aware agents. By building on these previous AI breakthroughs, AI systems can?unlock more real-world use-cases, providing experiences that are more personalised, transparent, and grounded in reality.

A System Architecture Design: Integrating Random Forest, Natural Language Processing, and Internet of Things to Predict Technical Carnapping in the Philippines

Jonell V. Ocampo, Jonathan M. Caballero — Fri, 18 Jul 2025 00:00:00 +0000

Technical carnapping, or rent-tangay, is a new deceptive scheme of stealing a car by virtue of a rental contract. Typical car anti-theft systems can only provide the location and status of the car. This means that by the time the vehicle is stolen, it is often too late. Even with the use of technology, this makes it hard for car owners and operators to prevent this new illegal scheme. This study features an architecture design for a car anti-theft system that integrates the use of natural language processing (NLP), random forest (RF) model, and internet of things (IoT) in predicting technical carnapping or rent-tangay in the Philippines. It highlights three major components, which are the black box or the hardware module, the mobile application, and the website application. The black box is responsible for gathering data inputs, including geographical location, recorded audio, and sensor outputs. The NLP pipeline is responsible for mining and processing text-based data from the audio recording. Whereas the RF model is responsible for scoring all of the inputs and using them to predict technical carnapping. The model developed in the study scored 100% in recall, 96.30% in accuracy, and 85.71% in F1 score. This implies the success and effectiveness of the design in predicting technical carnapping. This achievement significantly contributes to the body of work that focuses on developing security systems for cars, especially by effectively and efficiently implementing NLP and machine learning to the system. The study pushes the technological boundaries that can be explored in designing and developing car security systems.

Maturity in IT Monitoring: Enhancing Enterprise Preparedness for Critical Incidents

Manjunath Venkatram — Fri, 18 Jul 2025 00:00:00 +0000

In today's complex enterprise IT environments, the true measure of an organization's preparedness for critical incidents lies in the maturity of its IT monitoring capabilities. This maturity directly dictates how effectively IT teams can detect, navigate, and resolve incidents, ultimately minimizing downtime and business impact. High Mean Time To Detect (MTTD) and Mean Time To Resolve (MTTR) IT problems are directly linked to significant business losses, with IT downtime costing businesses over $100,000 per hour, and high-impact outages frequently exceeding $1 million per hour, sometimes lasting for days [5, 6, 7].

This white paper delves into the dual pillars of IT monitoring maturity: proactive monitoring with actionable alerting and comprehensive visibility for deep investigation and root cause analysis. We will explore how the proliferation of alert noise can severely impede incident triage, leading to significant delays and extended MTTD. A mature monitoring practice emphasizes the generation of critical, high-fidelity alerts that truly matter. Beyond alerts, effective incident response hinges on holistic visibility across all IT layers—network, application, infrastructure, end-user, and logs—ensuring real-time data capture and historical storage for context to drastically reduce MTTR.

Through a detailed use case of high CPU utilization on a server, we will illustrate the rigorous process of problem qualification and the multi-faceted investigation required to uncover root causes. This involves correlating data from diverse dependencies, from network traffic and application transactions to server health metrics and logs. The paper argues that true problem resolution aims for long-term fixes, moving beyond superficial adjustments to address underlying issues and build enduring IT resilience. Achieving IT monitoring maturity is not just about tools, but about establishing processes and data-driven insights that empower IT teams to fix problems faster and more effectively than ever before.

MCP Agents for Automated Cloud Compliance and Governance

Tue, 26 Aug 2025 00:00:00 +0000

Cloud computing has helped in reinventing the way businesses are run by offering scalable, flexible and cost-effective solutions. Yet, with more and more services on the cloud there is a difficulty of ensuring compliance and governance for these cloud environments that come from their complex, varied and dynamic nature of the clouds. Having manual, time-consuming, and error-prone traditional compliance management methods can speed up the process of compliance audits. In this paper, to resolve this problem, we implement Multi-Cloud Platform (MCP) agents using Artificial Intelligence(AI) for automatic cloud compliance and governance. MCP agents can monitor, analyze and enforce policies across multiple cloud environments to ensure compliance with industry standards, regulatory needs and internal governance principles. This system is recommended to be powered by AI, built on machine learning and natural language processing technologies which allows integrators to take better control of risk detection and mitigation. These agents can autonomously sift through massive amounts of cloud activity data, detect problematic configurations, and offer a real-time fix or suggestion Thus, it reduces manual interventions and brings a more effective, scalable and consistent set of cloud governance being enforced. This paper describes an umbrella architecture for enabling multiple compliance frameworks on MCP Agents. We also demonstrate how these agents provides cross cloud capability and can be controlled centrally with full visibility from a single dashboard. Using sophisticated, AI-driven models, these agents can predict potential compliance risks and prevent violations early on — enabling organizations to quickly identify security gaps before a breach or regulatory penalties arise. The simulation experiments confirm that our approach for AI driven MCP agents is faster and more accurate than traditional compliance checks. And, with AI plus MCP agents in the mix, it all adds up to a groundbreaking service that accelerates cloud compliance and empowers enterprises to easily strike out multi-cloud worlds armed with robust governance.

Revolutionizing Big Data: Scalable Pipelines and the Power of Data Lakehouse Architecture

Hasini Koka, Lahari Popuri, Devkiran Narayana, Jessica Harshad Patel — Tue, 26 Aug 2025 00:00:00 +0000

This paper studies how the Data Lakehouse architecture has the potential to change data analysis because it combines the most useful elements of data lakes and data warehouses into a single, scalable and cost-effective system. This looks at parts of the Lakehouse system, including open storage standards, extra data layers and engines that use ACID principles and points out why it is important to have scalable data pipelines. A comparison of warehouses, lakes and lake houses proves that lake houses are better equipped to handle different types of data tasks. By showing how finance, healthcare and retail use data lake houses to do complex analytics and machine learning with large data, this paper demonstrates how these systems enable organizations to avoid the common limits faced with traditional infrastructure. It also explains the tools and technologies involved in making Lakehouse work—for instance, Apache Spark, Delta Lake, Apache Airflow and Databricks and it looks at topics for further study, like live data, AI-friendly orchestration and data settings that are both safe and easy to use together. All of these insights explain how data pipelines and Lakehouse systems play an important role in the future of big data.

PolypNet: A Lightweight CNN Framework for Early Detection of Colorectal Polyps Using Deep Learning

Nishat Tasnim, Ahmed Mamun, Md. Imrul Kayes — Wed, 10 Sep 2025 00:00:00 +0000

Colorectal carcinoma is one of the most common reasons for carcinogenic death in the current world. Identifying the polyps that are present in the colon walls is one method to prevent this illness. However, a sparse number of research studies have been done to create a computer system that will detect the indisposition in the earlier stage. The enlargement of computer vision technology has accelerated the process by retrieving helpful information from the correlated data. Nonetheless, it is important to create an untrammelled system that will be able to sport colon polyps with better accuracy and training cost. In this research, we have delineated a Convolutional Neural Network (CNN) to emphasise Adenomatous, Hyperplastic and Serrated Lesions. The experiment of the network on the basis dataset has achieved an accuracy of 99.95% within a training time of only 18 minutes and 59 seconds. Stable learning efficiency was attained by the six-layer CNN with max-pooling and dropout regularisation.

A Unified Framework for Digital Delivery: Transition Strategies from Legacy to Cloud-Native Systems

Rajalingam Malaiyalan — Thu, 20 Feb 2025 00:00:00 +0000

In this essay, the paradigm changes in cloud migration strategies—from lift-and-shift to full cloud-native transformation—is examined. By examining technology elements, organisational factors, and architectural patterns, the paper offers a comprehensive framework for comprehending and deploying cloud native infrastructure. The demands of contemporary analytics workloads, real-time processing needs, and the exponential expansion of data volumes are becoming more and more difficult for legacy data warehouse systems to handle, despite their decades of dependability. Cloud-native tactics are revolutionising how businesses link heterogeneous systems, manage processes, and provide uniform user experiences. In order to address the particular security needs of cloud-native systems, this paper looks at a range of privacy-enhancing and trust-centric tools and strategies. In particular, a range of solutions are discussed, including cloud-native endpoint security solutions for guaranteeing trust and resilience in dynamic contexts, runtime protection platforms for real-time threat detection and responses, and service mesh technologies for secure service-to-service communication. To improve trust and transparency in cloud-native security, the significance of threat detection and response systems, cloud-native security information and event management (SIEM) solutions, and network security are also discussed. To guarantee comprehensive security in a cloud-native architecture, we also provide an extensive case study that illustrates how security measures are implemented across many levels, including application, network, infrastructure, security, and compliance. Organisations may strengthen the security posture of their cloud-native implementations by looking at these privacy-enhancing techniques and technologies. This will lower risks and guarantee the reliability of their data and apps in the dynamic ecosystem of today's digital world.