Volume 13 | Issue 10 |

 Abstract

A peer-to-peer (P2P) car-sharing service can be built using a decentralized approach, allowing users to interact directly and eliminating the need for a central authority. By leveraging smart contracts, the platform can automate transactions and agreements, ensuring both user privacy and fair pricing without relying on intermediaries. This system would also use a dedicated crypto token to facilitate direct payments between drivers and passengers, making the entire process more efficient.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

P2P car-sharing, Decentralized interaction, Smart contracts, Crypto token, Customer privacy, Fair pricing.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The early detection of emotional distress and self-awareness are crucial aspects of mental health that are often limited. This study introduces PersonaAI, a real-time AI- based digital twin framework designed to bridge this gap. By analyzing user responses to psychological questions and emotional inputs, PersonaAI creates a dynamic replica of a user's personality and emotional behavior. The system leverages a combination of a conversational model (GPT-4), a vector database (Pinecone/FAISS) for long- term memory, and a relational database (MongoDB/SQLite) for profile data. This approach allows the digital twin to generate behaviorally-aligned responses, enabling users to gain insights into their mental patterns, receive reflective prompts, and proactively manage their emotional well-being. This paper outlines the system's architecture, functional and nonfunctional requirements, and its potential to revolutionize personalized mental health support.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

AI, Digital Twin, Mental Health, Personality, Self-Awareness, GPT-4, Vector Database

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Managing finances can be daunting due to the complex financial landscape, lack of financial literacy, and difficulty tracking expenses or budgeting. Existing tools often need more personalization, rely on static budgeting, and provide generic investment advice. To address these limitations and enhance financial literacy and management, this paper proposes the development of an AI-powered personal finance assistant. The proposed assistant will utilize machine learning and natural language processing to provide a comprehensive financial overview, personalized insights and recommendations, and educational content tailored to users' needs. Key features include automated expense tracking, customized budgeting aligned with income and spending patterns, tailored investment advice based on risk appetite and goals, and proactive notifications about significant financial events. Specific metrics for evaluation will include improvements in financial literacy measured by pre-and post-use tests, quality of financial decision-making, user satisfaction scores, task completion rates, feature utilization, and user engagement levels. The expected outcomes include improved financial acumen, optimal financial decision-making, achievement of monetary goals, and secured financial futures for users.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Artificial Intelligence, Personal Finance, Financial Literacy, Web-based Applications, Financial Management

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Fake News , Text Classification , Machine Learning , NLP , Logistic Regression , SVM , TF-IDF

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Deepfakes have emerged as a significant threat in today's digital age, enabling the creation of highly realistic manipulated videos and images that are difficult to identify without special tools. These fake media can lead to issues like spreading false information, fraud, political manipulation, and reduced confidence in digital evidence. PixelTruth introduces an AI-driven forensic tool that uses advanced machine learning and deep learning techniques to detect and expose deepfakes. It uses CNN models, frequency analysis, and mixed feature extraction methods to spot inconsistencies in faces, lip movements, and texture patterns. This system can accurately and quickly detect deepfakes and is useful in journalism, law enforcement, and content moderation.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Deepfake, Forensic Analysis, Artificial Intelligence, Machine Learning, CNN, Digital Trust

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The hiring process is often time-consuming and inefficient, as recruiters must manually review large volumes of resumes to identify suitable candidates. This paper proposes an AI-powered Resume Matcher system that leverages Natural Language Processing (NLP) and Machine Learning (ML) techniques to automate resume-job description matching. The system extracts key features from resumes and job postings, applies vectorization methods such as TF-IDF and BERT embeddings, and employs matching algorithms to generate compatibility scores. Experimental results demonstrate that the proposed system achieves an accuracy of over 92% in candidate-job matching, significantly reducing recruitment time and improving hiring efficiency. This work highlights the potential of AI to transform recruitment by providing fair, scalable, and efficient solutions.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Resume Matching, Recruitment, Machine Learning, Natural Language Processing, TF-IDF, BERT

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The organization of college fests involves multiple activities such as event scheduling, participant registration, coordination among student committees, and communication with participants. Traditionally, these tasks are managed manually or through scattered digital tools, which often leads to inefficiencies, data inconsistencies, and communication gaps. To address these challenges, we have developed a College Fest Management Platform, a web-based application built using the MERN (MongoDB, Express.js, React.js, Node.js) stack. The platform offers secure authentication using JWT for administrators, enabling them to create, update, and delete events with ease. Students, on the other hand, are provided with a user-friendly interface to view and explore upcoming events in real time. The system not only reduces dependency on manual processes but also ensures data accuracy, transparency, and improved accessibility. By leveraging cloud-hosted services such as MongoDB Atlas for data storage and Vercel for deployment, the platform provides scalability and availability. Extensive testing was conducted to validate the system's performance, usability, and reliability, demonstrating its effectiveness in streamlining college fest management. This project highlights how integrating modern web technologies can transform traditional event management into a digitized, efficient, and interactive system, thereby enhancing the overall fest experience for both students and administrators.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

PixelTruth: College Fest Management

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The challenge of providing personalized education in modern classrooms has become increasingly complex due to diverse learning needs, varying cognitive abilities, and the growing demand for individualized instruction. Traditional educational approaches--from one-size-fits-all textbooks to static video content and limited adaptive learning platforms--are proving inadequate in addressing the unique learning pace and comprehension levels of individual students. These limitations not only hinder academic progress but also contribute to student disengagement, knowledge gaps, and reduced learning outcomes, particularly in foundational subjects during critical developmental years. Recent research has focused on intelligent tutoring systems that leverage artificial intelligence (AI), natural language processing, and adaptive content generation to overcome these educational challenges. Among these innovations, AI-powered video generation systems, similar to Google's NotebookLM approach, have emerged as promising solutions that can create personalized educational content while maintaining engagement and comprehension through dynamic visual and auditory elements. This survey compiles and examines advancements in AI-driven personalized video tutoring systems, with emphasis on student modeling, adaptive content generation, and real-time assessment integration. We analyze existing works that incorporate large language models (LLMs) and video generation technologies into educational frameworks, evaluate their effectiveness compared to traditional and hybrid learning approaches, and highlight their potential to reduce learning gaps, improve comprehension rates, and enhance overall educational outcomes. The study also identifies unresolved challenges including content accuracy verification, scalability across diverse curricula, real-time processing requirements for interactive questioning, and adaptation to varying technological infrastructure in educational institutions. This work provides a structured perspective on how AI-powered video tutoring systems can evolve within broader educational technology frameworks by synthesizing insights from current research trends in personalized learning, student assessment, and adaptive content delivery. The survey aims to serve as a foundational reference for future research, bridging AI-driven educational content generation with practical classroom applications for K-12 education.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Artificial intelligence, personalized learning, video-based tutoring, student modeling, adaptive content generation, intelligent tutoring systems, educational technology, K-12 education, interactive learning, deep knowledge tracing.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In this fast-paced world, the number of deaths due to accident is growing at an expeditious rate. Major reasons for these accidents are rash driving, drowsiness, drunken driving, carelessness, etc. An indicator of survival rates after detecting accidents is the time between the occurrence of accidents and the advent of medical care to the victim. The rapid growth of technology has made everything more facile and this advancement in technology additionally increased accidents. Due to this delayed medical attention, the accident victims might die as well. As a solution to these problems, we introduce a system that detects road accidents and will provide an alert message to the most proximate control room immediately. The camera module of the system is deployed in accident-prone areas. Whenever an accident occurs, it will detect the accident and immediately report about it to the nearby control room. The working of the system is based on deep learning techniques that use convolutional neural networks. By utilizing this system, many people can be saved from death.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Deep learning, image processing, neural networks

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The problem of managing traffic in contemporary cities has become more difficult due to the exponential increase in vehicle traffic and rapid urbanization. When it comes to managing dynamic and unpredictable road conditions, traditional methods--from manual police regulation to fixed-timer signal systems and limited sensor-based approaches--are becoming less and less effective. These inefficiencies exacerbate environmental issues by increasing fuel consumption and greenhouse gas emissions in addition to causing lengthy delays and driver stress. Recent studies have focused on intelligent traffic management systems that use computer vision, machine learning, and artificial intelligence (AI) to overcome these drawbacks. One of the most popular real-time object detection frameworks among them is You Only Look Once (YOLO), which provides excellent vehicle recognition accuracy and efficiency in a variety of traffic situations. With an emphasis on vehicle detection, density estimation, and adaptive signal control, this survey compiles and examines developments in YOLO-based traffic signal optimization. We examine previous works that incorporate YOLO into intelligent transportation systems, evaluate how well they perform in comparison to conventional and hybrid approaches, and emphasize how they can lower traffic, travel delays, and energy usage in general. The study also lists unresolved issues like robustness in inclement weather or low visibility, hardware constraints for real-time processing, and scalability to extensive road networks. This work offers an organized viewpoint on how YOLO-based systems can develop within larger smart city frameworks by incorporating insights from current research trends. The survey's ultimate goal is to act as a reference. point for further research, connecting computer vision methods powered by AI with practical intelligent traffic management applications.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Computer vision, traffic signal optimization, urban mobility, smart cities, artificial intelligence, YOLO, vehicle detection, traffic congestion, and intelligent traffic systems.

  License

Creative Commons Attribution 4.0 and The Open Definition