José Gabriel Carrasco Ramirez, CEO at Quarks Advantage, Jersey City, New Jersey. United States of America
The creation of proprietary frameworks for the development of neural models is essential to meet specific needs that generic frameworks cannot address. This article examines the key stages in the design of these frameworks and offers best practices for their effective implementation. It explores everything from needs identification and resource assessment to architectural design and implementation. Additionally, it emphasizes the importance of user-centered design and continuous evaluation to ensure the frameworks usability and adaptability to changing needs.
Proprietary frameworks, neural models, artificial intelligence, framework design, model optimization, user-centered design, continuous evaluation, scalability, performance optimization, data management, model training, regulatory compliance, explainable AI (XAI), agile methodology, security and privacy.
Kruti Shah, Preksha Patel, Varenya Uchil, and Prof. Pankaj Sonawane, SVKM‘s Dwarkadas J Sanghvi College of Engineering, Mumbai- 400056, India
The infestations of pests cause big losses in crops, both in quantity and quality. Present-day pest management systems are imprecise as they often rely on the traditional traps that give wrong data which can result in the incorrect estimates of infestation levels. Optimal trap placement is one of the requirements for traditional methods so as to enhance the accuracy of the monitoring data and hence cut down the overall costs. This chapter proposes the idea of “Ecobot: an IoT-based pest detection and eradication system” with ESP32-CAM for visual detection and infrared sensors for motion detection. It also includes INMP441 MEMS microphones for audio detection, alongside artificial neural networks for sound classification in ResNet to recognize images for accurate pest identification. Ecobot stands out due to its intelligent trapping system, which incorporates pheromone lures to attract pests effectively while minimizing pesticide use and handling large quantities without compromising trapping efficacy. Real-time alerts are sent across with the help of a Telegram bot to facilitate timely intervention with a substantial reduction in crop loss. The system can aid in improvement of accuracy in real-time monitoring and reduction in crop damage. It is well ahead of conventional methods in efficiency, scalability, and cost-effectiveness.
Sustainable, IOT, ANN, ResNet, ESP32, Pest Detection, Intelligent Trap System.
Nathan Andrie Ama, Bachelor of Science in Agribusiness, Southern Leyte State University, Hinunangan, Philippines
The integration of Internet of Things (IoT) technologies across industries has sparked significant innovation and entrepreneurial opportunities. Additionally, IoT has become a catalyst for innovation and entrepreneurship, yet existing studies often focus on its impact within single industries domains, leaving a gap in understanding its broader applications. This systematic review addresses gaps by identifying patterns and trend in IoT applications, innovation types, and entrepreneurial impacts across multiple industry domains including healthcare, manufacturing, healthcare, agriculture, retail, transportation, and energy. By synthesizing and categorizing findings from diverse studies, this study highlights the commonalities and distinctions in IoT enabled innovations and entrepreneurial activities. The study provides valuable insights into the diverse innovation and entrepreneurship, serving as a resource for researchers and practitioners seeking to explore IoT’s multifaceted potential.
Entrepreneurship, Systematic Review, Internet of Things (IoT), business, Industries.
Anastazja Drapata, University of Warsaw
AI is a technology of significant importance for use in the space and defense industry thanks to its features such as the ability to learn quickly and precision, which allows to achieve efficiency at a higher level than with the use of traditional weapons.The use of AI in the construction of space weapons is a controversial matter due to the choice of the level of regulation and the high risk of error, which may give rise to international liability.The increasing militarization of space raises ethical and legal doubts about the use of AI as a base technology for the production of space weapons, violating the purposes of using space set out in the 1967 Space Treaty, and also requires the adoption of legal regulations regulating its proper use in space research.Therefore, an analysis of AI applications in the production of space weapons was conducted.In order to reconstruct the legal norms governing the permissible use of AI, traditional legal inference, linguistic interpretation and the comparative method were used, referring to diplomatic documents and regulations issued by UN bodies.
AI, militarization of space, space, space weapons .
Meethun Panda1 and Soumyodeep Mukherjee2, 1Associate Partner, Bain & Company, Dubai, UAE, 2Associate Director, Genmab, Avenel - NJ, USA
This paper explores privacy and security frameworks tailored for Retrieval-Augmented Generation (RAG)-based Generative AI applications. These systems, while transformative in their capabilities, pose significant privacy and security risks. By leveraging advanced privacy-preserving techniques, robust governance frameworks, and innovative tools such as differential privacy and zero-trust architectures, this paper provides strategies for mitigating risks like data leakage, adversarial attacks, and compliance violations. Through theoretical and practical analysis, we present scalable approaches that align with global regulations such as GDPR and CCPA, ensuring operational performance and compliance.
Retrieval augmented generation, LLM, Privacy Preservation, Data Security, Adversarial Attacks, GDPR, CCPA, Differential Privacy, Governance, Secure AI Infrastructure, Data foundation.
Nidhi Joshi Parsai1 and Dr Sumit Jain2, 1Department of Information Science & Engineering, CMRIT Bangalore, India, 2Department of Computer Science & Engineering, Sage University, Indore, India
The rapid growth of Internet of Things (IoT) networking has created an urgent demand for strong protections to address new and sophisticated threats. Traditional Intrusion Detection Systems (IDS) struggle to manage the dynamic characteristics of IoT environments since they rely on static, signature-based methods. This research article presents a deep learning (DL)- based system that employs models such as Generative Adversarial Networks (GANs) and Convolutional Neural Networks (CNNs) to enhance IoT threat identification. This paper evaluates key DL algorithms, investigates scalability across different IoT domains, and addresses challenges such as resource limitations and data diversity. The proposed method enhances performance considerably, achieving a final detection accuracy of 95%, precision of 90%, and recall of 90% following 10,000 epochs of extensive training. In addition, it removes false positives and effectively addresses zero-day threats. The results emphasize the transformative capability of DL-enhanced IDS in safeguarding IoT networks with exceptional precision, immediate adjustment, and resilience against advancing cyber threats.
Intrusion Detection System (IDS), Deep Learning, Machine Learning, Cyber-attacks.
Akshaykumar Wankhade and Prema M. Diagavane, Research Scholar, Department of Electrical Engineering, GHRU, Amravati, (M.S.), India
In various applications, such as electric cars, energy storage devices, and portable gadgets, accurately measuring the current charge level is essential to safely administering and operating lithium-ion batteries. This investigation introduces an exhaustive audit of practical contemplates that have been led to validate the precision, dependability, and robustness of different SOC appraisal techniques under varying activity conditions, comprising altering charge/release rates, temperature profiles, and battery maturing impacts. The survey examines the execution of procedures, such as extended Kalman channels, particle channels, and adaptive onlookers, which have been embraced to address lithium-ion battery elements nonlinear and time-fluctuating nature. Additionally, the paper examines the criticalness of creating physics-based models that can get a handle on the underlying electrochemical cycles inside the battery and how these models can be incorporated with propelled appraisal calculations to improve SOC figure precision. The discoveries from this audit highlight the need for broad practical approval of SOC appraisal techniques in practical activity situations and the potential for additional upgrades through versatile and information-driven methodologies that can account for the unpredictable conduct of lithium-ion batteries.
State-of-charge appraisal, lithium-ion batteries, practical approval, activity conditions, versatile estimation, physics-based demonstrating.
Sai Ganesh Grandhi and Saeed Samet, School of Computer Science, University of Windsor, Windsor, Canada
Traditional authentication methods, such as passwords and biometrics, verify a user’s identity only at the start of a session, leaving systems vulnerable to session hijacking. Continuous authentication, however, ensures ongoing verification by monitoring user behavior. This study investigates the long-term feasibility of eye-tracking as a behavioral biometric for continuous authentication in virtual reality (VR) environments, using data from the GazebaseVR dataset. Our approach evaluates three architectures— Transformer Encoder, DenseNet, and XGBoost—on short- and long-term data to determine their efficacy in user identification tasks. Initial results indicate that both Transformer Encoder and DenseNet models achieve high accuracy rates of up to 97% in short-term settings, effectively capturing unique gaze patterns. However, when tested on data collected 26 months later, model accuracy declines significantly, with rates as low as 1.78% for some tasks. To address this, we propose periodic model updates incorporating recent data, restoring accuracy to over 95%. These findings highlight the adaptability required for gaze-based continuous authentication systems and underscore the need for model retraining to manage evolving user behavior. Our study provides insights into the efficacy and limitations of eye-tracking as a biometric for VR authentication, paving the way for adaptive, secure VR user experiences.
Continuous authentication, Virtual reality, Eye-tracking, Biometrics, Transformers.
Jon Laurence B. Wenceslao, Felicisimo V. Wenceslao, Jr., and Patrick D. Cerna, College of Information and Computing Studies, Northern Iloilo State University, Estancia, Iloilo, Philippines
This paper proposes a modified version of the Blowfish algorithm using permutation techniques. These techniques are key parts of the F-function of the algorithm. The first modification is in the bit operations while the second part employs the cyclical Shift Rows operation. We conducted simulations to test the security strength of the modified Blowfish algorithm in changing one bit from two input strings. The avalanche effect showed 54.687% and 53.125% changes in the output bits thus making our model secure. We also tested the execution performance of the original Blowfish algorithm and the modified Blowfish algorithm in the encryption and decryption processes. While the original version performs better than the modified version, no statistical difference was found between the two versions.
Blowfish Algorithm, Encryption, Cryptography, Permutation, Cyclical Shift Rows.
Mayank Singh, Abhijeet Kumar, Sasidhar Donaparthi, Gayatri Karambelkar, Fidelity Investments, Bangalore, Karnataka
The effectiveness of data catalogs hinges on the ease with which business users can look-up relevant content. Unfortunately, many data catalogs within organizations suffer from limited searchability due to inadequate metadata. This paper proposes a unique prompt enrichment idea of leveraging existing metadata content using retrieval based few-shot technique tied with generative large language models (Llama, GPT3.5). The literature also considers finetuning an LLM on existing content and studies the behavior of few-shot finetuned LLM (Llama2-7b) by evaluating their performance based on accuracy, factual grounding, and toxicity. The preliminary results exhibit more than 80% Rouge-1 F1 for the generated content. This implied 87%-88% of instances accepted as is or curated with minor edits by data stewards. By automatically generating descriptions for tables and columns in most accurate way, the research attempts to provide an overall framework to effectively scale metadata curation thereby vastly improving the data catalog searchability and adoption
Content Generation, NLG, Generative LLMs, Few-Shot Prompting, Data Catalog, Metadata Enrichment.
Mohammad Haseen ahmed, Department of English Language, King Abdul Aziz University, Jeddah, K.S.A
Data mining, also known as knowledge discovery in databases (KDD), involves the extraction of meaningful patterns, trends, and insights from large datasets using statistical and computational techniques. As the volume of data generated by organizations and individuals continues to grow, the significance of data mining has become increasingly apparent. This paper aims to provide a comprehensive overview of data mining applications across various fields, the strategies for harnessing its potential, and the challenges involved. Data mining has become an essential tool for extracting valuable insights from large datasets, with applications across various fields. By understanding the significance of data mining, its diverse applications, and the strategies for harnessing its potential, organizations can leverage data mining to drive decision-making, operational efficiency, and innovation. As data continues to grow in volume and complexity, the importance of data mining will only increase, making it a vital component of modern data-driven organizations.
operational efficiency, computational techniques, meaningful patterns, large datasets.
Copyright © DBDM 2025