Authors: Prof. Sangeeta Alagi, , Priti Jagdale, Swati More, Vaibhav Prasad

Abstract: The rapid advancement of deep learning technologies has enabled the creation of highly realistic synthetic media, commonly known as deepfakes. These manipulated videos pose serious threats to information integrity, personal privacy, national security, and public trust. This comprehensive literature survey examines the state-of-the-art approaches in deepfake detection, with particular emphasis on methods that combine Convolutional Neural Networks (CNNs) for spatial feature extraction with temporal analysis techniques. We systematically review detection methodologies, benchmark datasets, evaluation metrics, current challenges, and emerging research directions. This survey synthesizes findings from over 50 research papers published between 2018 and 2024, providing insights into the evolution of detection techniques and the ongoing arms race between deepfake generation and detection technologies.

Published by: vikaspatanker

Authors: Aakanksha Raghunath Chaudhari, Sharmistha Sujit Sarkar

Abstract: With the rapid growth of digital communication and online services, network security has become a primary concern for organizations and individuals. Traditional intrusion detection systems (IDS) rely heavily on predefined signatures, making them ineffective against zero-day attacks and unknown threats. To overcome these limitations, AI-based anomaly detection systems have emerged as a powerful approach for identifying unusual patterns in network traffic that may indicate malicious activity. This research introduces NetGuard, an intelligent system that leverages machine learning and deep learning techniques to detect anomalies in network traffic. The system provides real-time threat detection, reduces false alarms, and enhances network resilience against evolving cyber threats.

DOI: https://doi.org/10.5281/zenodo.17570618

 

Published by: vikaspatanker

Authors: Aryan Bhatt, Aryan Verma

Abstract: Artificial Intelligence (AI) methods, specifically Large Language Models (LLMs), are increasingly being employed by developers and students to produce source code. Though helpful, such AI-produced code is problematic in terms of plagiarism, originality, and academic honesty. Hence, differentiating between code written by humans and code generated by AI has become vital for the prevention of plagiarism. This article provides an empirical evaluation of current AI detection tools to determine how well they can detect AI-generated code in educational and coding environments. The findings indicate that most of the tools are ineffective and not generalizable enough to be useful for detecting plagiarism. In order to deal with this problem, we suggest a number of solutions, such as fine-tuning LLMs and machine learning-based classification based on static code metrics and code embeddings obtained from Abstract Syntax Trees (AST). Our top-performing model outperforms current detectors (e.g., GPTSniffer) and gets an F1 score of 82.55. In addition to that, we carry out an ablation study to study the contribution of different source code features to detection accuracy.

DOI:

 

Published by: vikaspatanker

Authors: Morziul Haque, 2Mohammed Shaik Fahad, Ansh Goyal, Bhagyashree .N. Singh, Priyanka Sahu, Dr. Basavaraj Neelur, Deepak Kumar Punna, Lavanya Dahiya

Abstract: In 1977, NASA launched two identical spacecrafts known as Voyager 1 and 2, which is the most important ambassador of mankind to the universe. Voyager as a project which was originally intended to be a planetary exploration mission, transformed into a historic project which incorporated both scientific, engineering and humanistic goals. Throughout a period of close to five decades, the Voyagers have unrelentingly provided deliveries in terms of firsts in regard to the outer planets, the heliosphere and the interstellar medium. They are nowadays the well-known stepping-stones of human inquisitiveness and venture beyond the solar frontier. The theory discussed in this paper is a literature review about the current scholarly debate around the issue of the scientific success, the engineering strength, and cultural meaning of the Voyager mission in terms of the 21st century digital era. It also calls attention to modern reinterpretations of Voyager data with these aspects may involve the use of artificial intelligence (AI) and astrophysical modeling, as well as the continued debate surrounding the following, interstellar communication and preservation. Through the lens of both the empirical heritage and with an emphasis on the philosophical influence of the Voyager program, this review explores the mission against the background of 21st century space exploration and human self-understanding.

DOI: https://doi.org/10.5281/zenodo.17568692

 

Published by: vikaspatanker

Authors: Garima, Ravi Kumar Rana, Niranjan Singh Rathee

Abstract: In the current research work a new Schiff base (2,2'-((1E,1'E)-((4-methyl-1,2-phenylene)bis(azaneylylidene))bis(methaneylylidene))bis(4-bromophenol) (H2L) and its metal complexes were prepared using condensation reaction of 3,4-diaminotoluene and 5-bromo salicylaldehyde. The Schiff base was further coordinated with Co2+, Ni2+, Cu2+& Zn2+ metal ions to synthesize its 1:1 metal complexes. All the synthesized compounds were examined using a variety of characterisation methods, including proton-NMR, electronic, mass, ESR, IR spectroscopy and TGA. FT-IR and NMR spectral data elucidate that the metal ions are coordinated with the tetradentate ligand through 2-N (imine) and 2-O (hydroxyl) atoms. All of the metal complexes were assumed to have an octahedral geometry based on the UV-Visible spectra. Antibacterial & antifungal activity of synthesized product was tested. The results demonstrated that all the sample exhibited considerable antimicrobial properties.

DOI: https://doi.org/10.5281/zenodo.17557767

 

Published by: vikaspatanker

Authors: Praniti Gijare, Sneha Kunnummal, Suhani Heblikar, Harsh Sakhare, Rushabh Parab, Reshma Sonar

Abstract: Phishing attacks targeting college students have surged by 224% in the education sector during 2024, In recent months, attacks aimed at stealing login details have exploded in volume, with credential-related phishing growing at an unprecedented rate and now representing the fastest-rising threat faced by campus communities. Methods that rely mainly on lectures or passive training have not made a substantial impact on how well students identify or avoid phishing threats, leaving many learners at risk despite completing such programs in reducing phishing susceptibility, with studies revealing minimal behavioural change despite widespread implementation. This research investigates whether interactive coding simulations using Python-based phishing detection exercises can significantly improve college students' ability to identify and avoid phishing attacks compared to conventional lecture-based training. A quasi-experimental pre-test post-test design employed 90 undergraduate students across three groups: interactive simulation training (n=30), traditional lecture-based training (n=30), and control group (n=30). The interactive group developed basic Python scripts to detect phishing characteristics including suspicious URLs, sender anomalies, and social engineering tactics. Results indicate the interactive simulation group demonstrated Students who took part in coding-based, hands-on exercises were able to spot phishing attempts nearly twice as well as those who received traditional classes, showing a remarkable 42% boost in detection skills over standard methods 18% in the lecture-based group and The students who didn’t receive any security Students who didn’t participate in any cybersecurity activities barely improved at all, showing almost no change in their ability to recognize phishing scams; this highlights that without fresh skills, people generally stick to old habits even when digital threats are increasing 5% improvement, which suggests that without any active intervention, most people simply continue their usual habits even if they face ongoing cyber risks. The findings suggest hands-on coding simulations provide superior learning outcomes through experiential engagement, addressing a There is a clear need for practical, engaging Cybersecurity education still relies heavily on traditional classroom approaches, most of these traditional methods don’t really equip students for the types of scams and online risks they will actually encounter in their daily lives, leaving important gaps in both confidence and readiness the constantly changing landscape of digital threats fail to prepare students for real-world online risks they often lack the tools and confidence, most students aren’t equipped with the practical skills they need to spot and steer clear of today’s online dangers, which means entire groups remain vulnerable unless more effective and engaging education is provided.

Published by: vikaspatanker

Authors: C.Jaya Prakash Reddy, R.Jaswanth, K.Rajeshkumar

Abstract: In a world where digital threats are on the rise, especially in education, we designed a mobile-first LMS (Learning Management System) to promote cybersecurity awareness in universities. Using Flutter for cross-platform app development and Firebase for cloud backend, this solution helps students and staff learn, interact, and stay informed—even offline. Key features include video modules, real-time quizzes, secure authentication, and user-friendly dashboards tailored for students, instructors, and admins. It’s lightweight, fast, scalable—and ready to make cybersecurity education smarter and more accessible.

Published by: vikaspatanker

Authors: Ms.Sabitha K, Bhavayazhinitha S V, Jashwanth M U, Gunal S, Gokulnath K

Abstract: Organizations of the modern digital era create massive volumes of operation and security event data that must be monitored in real time and responded to immediately. Server-based monitoring systems are generally connotated with bad scalability, high overhead maintenance, and delayed response to alerts. To solve all these problems, SeclogX – Serverless Real-Time Monitoring and Alerting System is proposed, using Amazon Web Services (AWS) to create an entirely serverless, event-driven system. The system makes use of various AWS services, including Amazon API Gateway (REST & WebSocket), AWS Lambda, Amazon DynamoDB, Amazon SNS, Amazon CloudWatch, and Amazon S3 with CloudFront, in order to provide high availability, low latency, and real-time processing. Frontend dashboard ran over the internet on Amazon S3 and CloudFront and admin- accessible site-deployed web site and dashboard used for real-time system status and live event visualization. Anomaly detection is done automatically, data processing is done through Lambda functions, event logs are stored in DynamoDB, and alerts are triggered through SNS. Monitoring and logging are provided through CloudWatch for system health and operational intelligence. With the implementation of this serverless architecture, SeclogX removes the maintenance overhead, enables scalability, and conserves the cost while not sacrificing its real-time alert on crucial events. The result indicates that the model is an affordable, secure, and scalable event monitoring system that can be adopted by various industries.

DOI: https://doi.org/10.5281/zenodo.17551088

 

Published by: vikaspatanker

Authors: Keerti Pal, Prof. Jayshree Boaddh, Prof. Rahul Patidar

Abstract: Product Sales Dataset is a comprehensive collection of sales data for a wide range of products available on the E-commerce e-commerce platform. This kind of dataset provides invaluable insights into customer behavior, product performance, and market trends, making it an essential resource for data analysis, market research, and business strategy development. This dataset is indispensable for market research, allowing businesses to discern market trends, consumer preferences, and competitive landscapes. This paper presents a comprehensive approach to customer behavior analysis and predictive modelling within the context of supermarket retail. This paper finds techniques that extract patterns in shopping data for the learning and prediction of user preference. This work list different proposed models with techniques. Paper has list various evaluation parameters of user purchase prediction models.

DOI: http://doi.org/10.5281/zenodo.17542512

Published by: vikaspatanker

Authors: Nweke James, Emenike Wami, Awajiogak Anthony Ujile, T.O. Goodhead

Abstract: This study evaluates biodiesel production from water hyacinth (WH) via transesterification, highlighting its potential as a sustainable renewable energy source. Lipids were extracted from WH using Soxhlet and maceration methods, yielding modest oil content. Five methanol-to-oil molar ratios (4:1, 5:1, 6:1, 7:1, 8:1) were tested, with the 6:1 ratio in combination with a NaOH catalyst producing the highest biodiesel yield of 88.21%. The biodiesel obtained exhibited a cetane number of 57.66, meeting ASTM D6751 standards and indicating excellent ignition quality suitable for high-efficiency diesel engines. Kinetic modelling. of the transesterification reaction was conducted to determine rate constants and conversion efficiencies, providing critical data for process optimization and scale-up. Using Python 3.11 with the Levenberg–Marquardt algorithm, the kinetic model closely fitted the experimental data, enabling accurate prediction of reaction progress and substrate conversion. These results demonstrate that water hyacinth is a viable feedstock for biodiesel production, offering both energy recovery and environmental management benefits. The study provides validated operational parameters and kinetic insights for the development of cost-effective, scalable biofuel production from aquatic biomass.

DOI: https://doi.org/10.5281/zenodo.17542202

 

Published by: vikaspatanker