Authors: Siddhi Ekawade, Apurva Jate, Arya Kamble, Sharvari Kate, Prof.Pradnya Satpute

Abstract: Artificial Intelligence has made it easy to create realistic images, videos, and texts. These technologies have been misused to create deepfakes and online scams, which can lead to the spread of misinformation, financial scams, and cybersecurity attacks. It is hard to detect such content by human beings, as it is time-consuming. Hence, there is a need to develop an automated detection system for AI-generated content. The proposed project aims to develop a multimodal AI-generated content detection system that can analyze images, videos, and texts to detect potentially fake or scam content. The system can detect deepfakes in images and videos using a Convolutional Neural Network (CNN) model, and it can also analyze the text messages sent by the user to detect scams using a machine learning-based approach. The application has been developed as a web application using the Flask framework in Python. This processed media is analyzed, and the important features are identified, providing a probability score on whether the media is real or fake. The output is given in percentage probability, making it easier for the user to interpret the results. All analysis results are stored in a SQLite database, which is used for monitoring and administrative purposes. This proposed system has shown how deep learning and machine learning can be combined into a single framework to detect AI-generated content. This type of system can be used to enhance digital security, helping users identify fake media and possibly scam messages.

 

 

Published by: vikaspatanker

Authors: Megha Garud, Lalit Gaikwad, Prakash Mane, Ranjit Misal, Amey Phatak, Premraj Takawale

Abstract: The increasing adoption of Electric Vehicles (EVs) has brought significant attention to the availability and efficiency of charging infrastructure. Although governments and private organizations are actively deploying public charging stations, their limited number and uneven distribution continue to pose challenges for EV users. In many cases, users experience long waiting times or difficulty locating nearby charging facilities. This survey paper presents an EV Charger Sharing Platform that encourages the utilization of privately owned EV chargers through a web-based system. The platform enables charger owners to list their chargers and EV users to search, view, and book available charging slots based on location and availability. Developed using standard web technologies such as HTML, CSS, and JavaScript, the system aims to improve charger accessibility, reduce waiting time, and promote sustainable transportation. By adopting a sharing-economy approach, the proposed solution offers a cost-effective and scalable alternative to traditional public charging infrastructure.

Published by: vikaspatanker

Authors: Krishna Prisad Bajgai, Dr. Saroj Khanal, Dr. Bhoj Raj Ghimire

Abstract: Type 2 Diabetes Mellitus (T2DM) remains a major global health burden and a leading contributor to cardiovascular, renal, and hepatic mortality. Traditional risk assessment models rely on limited clinical parameters and fail to capture complex nonlinear interactions among multi-organ biomarkers. This study proposes a comprehensive machine learning (ML) and deep learning (DL)-based survival modeling framework to predict mortality risk in T2DM patients using multi-organ biomarkers, including fasting blood glucose, HbA1c, serum creatinine, triglycerides (TG), total cholesterol, LDL, HDL, liver enzymes (ALT, AST), and fatty liver indicators. Using the National Health and Nutrition Examination Survey (NHANES) linked mortality dataset, we compare Cox Proportional Hazards, Random Survival Forest (RSF), Gradient Boosting Survival (GBM), DeepSurv, and Long Short-Term Memory (LSTM) models. Performance was evaluated using Concordance Index (C-index), time-dependent Area Under Curve (AUC), Hazard Ratio (HR), and Brier score. Results show DeepSurv achieved the highest C-index (0.82), followed by RSF (0.79), outperforming traditional Cox regression (0.72). SHAP-based feature importance revealed HbA1c, creatinine, triglycerides, and ALT as dominant mortality predictors. Risk stratification analysis demonstrated clear separation between low-, medium-, and high-risk groups (log-rank p < 0.001). The findings highlight the superiority of nonlinear survival models for mortality prediction in T2DM and provide clinically interpretable insights for personalized risk management.

DOI: https://doi.org/10.5281/zenodo.18886211

Published by: vikaspatanker

Authors: Deepak Tomar, Kismat Chhillar, Dhruv Srivastava

Abstract: This paper examines the potential of blockchain technology in strengthening of academic data integrity within institutions of higher education by addressing persistent challenges of credential fraud, limited traceability, record falsification and fragmented oversight in traditional centralized systems. Grounded in contemporary research on decentralized architectures and verifiable credentials, the study analyzes weaknesses in existing platforms of student management and proposes a conceptual model that is integrating key blockchain principles of distributed consensus, immutability and smart contracts, with requirements of integrity such as auditability, verifiability, non-repudiation and selective disclosure. The proposed model provides an outline of a consortium-based platform for management of transcripts, qualifications, assessments and co-curricular records, designed to interoperate with existing databases of institutions while complying with regulations of data protection and sovereignty regulations such as FERPA and GDPR. Scenario-based evaluations suggest improvements in verification efficiency, inter-institutional trust and provenance tracking, along with reductions in administrative overhead and faster dissemination of academic records to external stakeholders. The study also critically considers practical challenges that are related to jurisdictional interoperability, scalability, institutional resistance and governance.

 

 

Published by: vikaspatanker

Authors: Dr. L. N. Malviya

Abstract: To improve the main shortcomings of insufficient nutrients, high salinity and low productivity of soils, soil testing and analysis are crucial processes. Understanding the composition, quality, and characteristics of soil is essential for making informed decisions related to land use, crop management, building foundation design, and environmental remediation. Soil testing and analysis is very important in agriculture, construction, environmental science, and various other fields.

 

 

Published by: vikaspatanker

Authors: Akhilash Pennam

Abstract: This paper proposes an Artificial Intelligence (AI)–enabled cloud-based CRM architecture developed on the Salesforce platform to modernize gas station operations through intelligent automation and predictive analytics. The system integrates fuel booking, inventory management, supplier coordination, and customer interaction into a unified digital platform. AI techniques including time-series forecasting, anomaly detection, and customer behavior analytics are embedded to transform operational data into predictive insights. Machine learning models analyze historical transactions to forecast fuel demand, optimize inventory levels, and detect abnormal operational patterns. Salesforce automation tools such as Flows and Apex triggers enforce business rules, while AI-driven dashboards provide real-time predictive decision support. Experimental evaluation demonstrates improved forecasting accuracy, reduced operational errors, faster transaction processing, and enhanced managerial decision-making. The proposed architecture demonstrates how AI can elevate traditional CRM systems into intelligent, scalable, and proactive operational platforms suitable for multi-branch fuel retail environments.

DOI:

 

 

Published by: vikaspatanker

Authors: Ms. Dipti Rathod

Abstract: Artificial Intelligence (AI) has significantly transformed the field of computer vision by enabling machines to interpret and analyze visual data with high accuracy and efficiency. Computer vision, a major branch of AI, focuses on developing systems that can acquire, process, and understand images and videos in a manner similar to human vision. With the advancement of machine learning and deep learning techniques—particularly Convolutional Neural Networks (CNNs)—computer vision systems have achieved remarkable improvements in tasks such as image classification, object detection, facial recognition, and image segmentation. The integration of AI into computer vision has led to widespread applications across various industries, including healthcare, autonomous transportation, security and surveillance, retail, agriculture, and manufacturing. In healthcare, AI assists in medical image analysis and disease detection; in transportation, it powers self-driving vehicles; and in industrial sectors, it enhances quality inspection and automation. Despite these advancements, challenges such as the need for large labeled datasets, high computational costs, security risks, bias in algorithms, and ethical concerns remain significant issues in the computer industry. This research examines the role of artificial intelligence in the development of computer vision technology, explores its major applications, and highlights the key problems that need to be addressed. The study concludes that while AI-driven computer vision has revolutionized modern computing, continued research, ethical governance, and technological innovation are essential to fully realize its potential and ensure responsible implementation.

 

 

Published by: vikaspatanker

Authors: Nagaraju Goshikonda

Abstract: The digitalization of critical infrastructure sectors of the U.S. economy such as energy, transportation, healthcare, and defense has expanded the cyber-attack surface at a rapid pace. The old models of perimeter-based security are no longer effective against complex attacks, like advanced persistent attacks (APTs), insider attacks and AI-assisted cyber-attacks. This paper will suggest AI-based Zero Trust Security Architecture (ZTSA) adapted to secure the critical infrastructure in the United States. The framework incorporates behavioral analytics, federated learning, and adaptive risk scoring, that allow one to continue verification and intelligent response to threats. The predictive and generative AI models are utilized to simulate the attack scenario, improve anomaly detection, and automate policy enforcement. Experimental assessment based on simulated critical infrastructure datasets is shown to have a higher detection rate of 95.8 and a 30% lower rate of false positives than traditional zero-trust systems. The outcomes show that AI-enhanced zero-trust models will be capable of mitigating critical infrastructure in the US to a considerably greater extent in terms of resilience, scalability, and mitigation of threats in real-time.

DOI:

 

 

Published by: vikaspatanker

Authors: Dipak Kadve, Vaishali Suryawanshi, Aditi Choure, Pratibha Ghodake

Abstract: In recent years, digital streaming platforms have become an integral part of students’ daily routines. Among various forms of online entertainment, anime has gained significant popularity among college students in urban Indian cities, including Pune. While entertainment media can provide relaxation and emotional engagement, concerns are often raised regarding excessive viewing habits and their potential academic implications. The present study explores the relationship between anime consumption patterns and academic productivity among college students in Pune. A structured online questionnaire was administered to 150 respondents across different academic disciplines. The study examined variables such as daily viewing duration, binge-watching behavior, late-night streaming habits, sleep duration, study hours, and self-reported academic performance. The findings suggest that moderate anime consumption does not significantly affect academic productivity. However, extended late-night viewing and frequent binge-watching were associated with reduced study hours and irregular sleep patterns. The study highlights the importance of balanced digital engagement and responsible time management among students.

 

 

Published by: vikaspatanker

Authors: Dr. Snehal Godse, Mr. Prathameshsingh U. Rajput, Prof. Apurva Shende

Abstract: Artificial Intelligence (AI) has evolved from task specific systems to increasingly adaptive and socially responsive agents. As research advances toward Artificial General Intelligence (AGI), a transformative shift is anticipated not only in computational capability but also in human–machine relationships. This study explores the dual nature of AGI—conceptualized as “Syrup and Sword”—wherein emotionally intelligent systems may deepen human closeness while simultaneously introducing existential and psychological risks. Drawing upon attachment theory, alignment research, and human–computer interaction (HCI) scholarship, this paper develops an integrative conceptual framework linking AI capability progression with attachment intensity and societal outcomes. A mixed-method research design is proposed, combining qualitative thematic analysis and quantitative experimental surveys to examine emotional scaling from Narrow AI to AGI-level systems. The study identifies key drivers of attachment such as empathy simulation, memory continuity, adaptive responsiveness, and perceived moral agency. It further analyzes risks including emotional dependency, anthropomorphic projection, manipulation, and value misalignment. The proposed “Syrup vs. Sword Framework” offers a structured lens to evaluate how increasing cognitive and affective sophistication in AGI could produce both enhanced well-being and destabilizing consequences. The findings contribute to interdisciplinary discourse by bridging psychological, ethical, and technical perspectives, emphasizing the necessity of emotionally aware governance in future AGI development. Index Terms—Artificial General Intelligence (AGI), Human–AI Attachment, AI Ethics, Alignment Problem, Emotional AI, Existential Risk, Human–Computer Interaction.

 

 

Published by: vikaspatanker