Category Archives: Uncategorized

Authors: Zaibindah Rafeeq Pandit, Sabit Aslam, Rehana Jan, Irfan Rasool

Abstract: Conversational agents, or more popularly called virtual assistants or chatbots, are now a unifying interface for modern digital ecosystems, enabling seamless human-computer interaction. Fueled by unprecedented accelerations in Artificial Intelligence (AI), Machine Learning (ML), and Natural Language Processing (NLP), these agents evolved from script-based rules to sentient agents with the capability to understand context, sentiment, and intent. Transformer-based models such as GPT and BERT have significantly improved fluency, coherence, and chatbot response flexibility so that conversations could be more human-like. The present paper follows the historical progression of conversational agents from the initial symbolic systems such as ELIZA to modern-day deep learning models. It covers significant architectural components like intent recognition, conversation management, and response generation with emphasis placed on the intersection of speech-to-text (STT) and text-to-speech (TTS) for voice interaction. The book also looks into popular frameworks and toolkits used to develop and deploy chatbots into real-world applications across healthcare, education, customer support, and mental health. Moreover, the paper highlights major challenges hindering the robustness of current systems, including data bias, hallucination, context limitations, and lack of emotional intelligence. Moral implications—particularly of fairness, privacy, and explainability—are argued against in terms of novel guidelines and mitigation strategies. A modular, LLM-assisted architecture is suggested to demonstrate practical implementation with inherent evaluation metrics. Finally, the paper outlines guidance for subsequent research and development, calling for emotionally smart, multi-lingual, and culturally sensitive conversational agents that are ethics-compliant and highly accessible and performing.

DOI: https://doi.org/10.5281/zenodo.18739755

 

Published by: vikaspatanker

Authors: Durgesh Nishad, Purvesh Patil, Vedant Chaudhary, Ghansham Bordekar, Dr.Umesh Pawar

Abstract: The MealSphere Management System is a modular, automated platform designed to streamline food ordering, inventory tracking, billing, and administrative operations. Traditional restaurant workflows reliant on handwritten logs and disconnected tools often result in delays, errors, and poor visibility. MealSphere resolves these inefficiencies through a centralized system that integrates authentication, menu management, inventory deduction, digital billing, and analytics. It supports both offline and online modes, ensuring operational continuity and real-time synchronization. This paper presents the system’ s architecture, implementation, and performance evaluation, demonstrating its scalability across restaurants, hostels, cloud kitchens, and canteens.

Published by: vikaspatanker

Authors: Apurva Tanaji Bhosale, Gaurav Prabhakar Pandhare, Suman Ravi Rathod, Nikhil Megharaj Tikande

Abstract: The Fully Automatic Solar Grass Cutter is an innovative, eco-friendly solution designed to automate lawn maintenance while utilizing renewable energy. This system operates entirely on solar power, eliminating the need for conventional fuel or external electrical supply. The solar panel mounted on the device captures sunlight and converts it into electrical energy, which is stored in a rechargeable battery to power the DC motors and control unit.The grass cutter is equipped with automated navigation and obstacle detection mechanisms using sensors, enabling it to move independently across the lawn while avoiding collisions. A microcontroller is used to control the movement of the wheels and the cutting blade motor, ensuring efficient and uniform grass trimming. The automation reduces human effort, operational cost, and environmental pollution compared to traditional petrol-powered grass cutters.This project emphasizes sustainability, energy efficiency, and smart automation. It is suitable for residential lawns, gardens, parks, and institutional grounds. By integrating renewable energy with robotic technology, the Fully Automatic Solar Grass Cutter provides a cost-effective, low-maintenance, and environmentally friendly alternative for modern lawn care applications.energy, which is stored in a rechargeable battery. This stored energy powers the DC motors responsible for blade rotation and vehicle movement. A microcontroller-based control unit manages the overall operation of the system.The machine is equipped with sensors for obstacle detection and autonomous navigation, enabling it to operate without human intervention. The automatic mechanism ensures uniform grass cutting while reducing manual labor, fuel consumption, and environmental pollution. Compared to conventional petrol-driven grass cutters, this system offers low operational cost, minimal maintenance, and zero carbon emissions.

DOI:

 

 

Published by: vikaspatanker

Authors: Apurva Tanaji Bhosale, Gaurav Prabhakar Pandhare, Suman Ravi Rathod, Nikhil Megharaj Tikande

Abstract: The Fully Automatic Solar Grass Cutter is an innovative, eco-friendly solution designed to automate lawn maintenance while utilizing renewable energy. This system operates entirely on solar power, eliminating the need for conventional fuel or external electrical supply. The solar panel mounted on the device captures sunlight and converts it into electrical energy, which is stored in a rechargeable battery to power the DC motors and control unit.The grass cutter is equipped with automated navigation and obstacle detection mechanisms using sensors, enabling it to move independently across the lawn while avoiding collisions. A microcontroller is used to control the movement of the wheels and the cutting blade motor, ensuring efficient and uniform grass trimming. The automation reduces human effort, operational cost, and environmental pollution compared to traditional petrol-powered grass cutters.This project emphasizes sustainability, energy efficiency, and smart automation. It is suitable for residential lawns, gardens, parks, and institutional grounds. By integrating renewable energy with robotic technology, the Fully Automatic Solar Grass Cutter provides a cost-effective, low-maintenance, and environmentally friendly alternative for modern lawn care applications.energy, which is stored in a rechargeable battery. This stored energy powers the DC motors responsible for blade rotation and vehicle movement. A microcontroller-based control unit manages the overall operation of the system.The machine is equipped with sensors for obstacle detection and autonomous navigation, enabling it to operate without human intervention. The automatic mechanism ensures uniform grass cutting while reducing manual labor, fuel consumption, and environmental pollution. Compared to conventional petrol-driven grass cutters, this system offers low operational cost, minimal maintenance, and zero carbon emissions.

DOI:

 

 

Published by: vikaspatanker

Authors: Rupal Bula, Dr. Vandan Tewari, Mrs Sonu Airen

Abstract: The recommendation system is a classification of machine learning [1] that uses features to help anticipate, compact and find what people are looking for an exponentially growing number of options. It is an artificial intelligence algorithm in machine learning, which uses big data to recommend more related items to consumers. These can be based on different criteria, which includes past purchase, search history, demographic information and other factors. These systems are designed to predict what a user might like based on various factors. They are extensively used in various domains including e-commerce, streaming services, social networks and content platforms. In this paper we have proposed a novel movie recommendation system that effectively addresses the cold start user problem by leveraging sequence modeling techniques [2]. Traditional recommendation systems struggle with new users due to the lack of historical interaction data. Our approach utilizes sequence modeling, specifically Long Short Term Memory (LSTM) networks, which predict user preferences based on initial interactions. By analyzing the sequence of the movie watched, our model can generate accurate recommendations even with minimal user data.

DOI: https://doi.org/10.5281/zenodo.18723361

Published by: vikaspatanker

Authors: S.Rutuja, P. Sanskruti, P. Shreya, B.Nandini, M.Priya

Abstract: The core aim of the given project is to create an Web application named MediMinder, that assists people in keeping track of their daily health routines without much effort. The application gives smart reminders to take medicine, attend the checkups and make appointments with the doctor.The machine learning (ML) will learn the habits of the user and get their schedules in order by creating personalized and voice-activated reminders that include particular date and time, and medicine information.With the application, all users can enjoy better health management and safety by receiving personalized, voice-based, and specifically set reminders with a specific date, time, and specific medicine information.

Published by: vikaspatanker

Authors: Urgen Maharjan, Gakegni Ngaste Aperhy, Sahasra Nagaraj, Sahil Gupta, Prithi M

Abstract: The recent fast digitalization of the educational system has seen the introduction of online examination systems in universities, examining bodies and professional organizations in large numbers. Even though these platforms enhance accessibility, scalability, and operational efficiency, they present vital security and integrity issues that pose a threat t o the credibility of online tests. The significant security issues are impersonation by candidates, contract cheating, leaking of question papers, tampering with answer scripts, unauthorised collusion, malware-based attacks, network breach and massive distributed denial-of-service (DDoS) attacks. Also, the introduction of remote proctoring systems based on AI provokes serious concerns about privacy, ethical and data protection issues. The currently in place security systems like password-authenticated access and crude browser limitation are inadequate in stopping advanced methods of cheating as well as online threats. The modern studies suggest sophisticated measures such as multi-factor authentication, facial recognition-based and key-stroke dynamic-based biometric verification, end- to-end encryption algorithms to ensure safe data delivery, blockchain-based transparent record keeping, behavioral analytics, abnormality detection approaches that use AIs, and lockdown browsers. Nevertheless, such solutions tend to cover single facets of security and can pose issues of computational cost, scalability, implementation cost and user confidentiality. The paper provides the in-depth examination of security issues with online examination systems and reviews the modern research-based solutions in a systematic way. The comparative assessment framework is created to determine the efficiency, practicability and constraints of different security mechanisms. According to the results, a multi-layered security architecture will be introduced incorporating the elements of authentication, communication security, intelligent monitoring, and tamper-resistant data management to guarantee the examination integrity, confidentiality, availability, and accountability. The research concludes that a combination of defense in depth strategy is crucial towards the attainment of secure, scalable and reliable online examination ecosystems.

Published by: vikaspatanker

Authors: Shantanu Shahaji Gaikwad, Aryan Dattatray Vibhute, Sharvari Arun Shedage, Yogita Sunil Kamble

Abstract: In today’з modern and faзt-moving world, the demand for зmart and automated зolutionз haз increaзed зignificantly, eзpecially in the field of pet care. Thiз reзearch focuзeз on the development of an automatic pet food diзpenзer deзigned to provide automated, preciзe, and intelligent feeding зolutionз to зupport pet ownerз with buзy lifeзtyleз. Many pet ownerз face challengeз in maintaining conзiзtent feeding зcheduleз due to work commitmentз, travel, and other daily reзponзibilitieз. The propoзed зyзtem aimз to зolve thiз problem by enзuring that petз receive timely and accurate mealз without conзtant human зuperviзion. Theзe зyзtemз integrate microcontrollerз, зenзorз, real-time clock moduleз, and Internet of Thingз (IoT) technologieз to enable зcheduled feeding, accurate portion control, and remote monitoring through mobile applicationз. The microcontroller actз aз the central control unit, coordinating the timing mechaniзm and diзpenзing proceзз. Senзorз are uзed to detect food levelз, monitor bowl зtatuз, and enзure proper food diзtribution. Through IoT connectivity, uзerз can acceзз the зyзtem remotely, adjuзt feeding зcheduleз, control portion зizeз, and receive real- time notificationз regarding feeding activitieз and зyзtem performance. Advanced modelз incorporate data tracking and intelligent featureз that record feeding hiзtory, analyze conзumption patternз, and provide inзightз to improve feeding accuracy and maintain optimal pet health. Some зyзtemз may alзo include voice recording featureз, camera monitoring, and emergency alertз to further enhance functionality and зecurity. Theзe зmart featureз help prevent overfeeding and underfeeding, reduce food waзtage, and зupport balanced nutrition. Overall, automatic pet food diзpenзerз enhance convenience, conзiзtency, and reliability in daily feeding routineз. They contribute зignificantly to improved animal welfare by enзuring petз receive proper nutrition at the right time. At the зame time, they reduce human effort, minimize зtreзз for pet ownerз, and repreзent an important зtep toward the integration of зmart technology in modern pet care management.Top of Form Bottom of Form

DOI: https://doi.org/10.5281/zenodo.18721934

Published by: vikaspatanker

Authors: Ms. Meenakshi, Dr. Shweta Mishra

Abstract: Regression testing plays a critical role in ensuring the continued functionality and quality of software as changes, updates, or bug fixes are implemented. Given the complexity and the growing size of modern software systems, managing large test suites efficiently is becoming increasingly difficult. This paper focuses on the importance of test case prioritization, a technique aimed at ordering test cases in a way that maximizes fault detection early while minimizing resource utilization. The study compares various theoretical models and methods of test case prioritization, including random, requirement-based, code-based, and model-based prioritization. It highlights the benefits of prioritization in improving test efficiency and the challenges associated with its implementation. Additionally, the paper reviews both traditional and advanced prioritization techniques, including risk-based prioritization and AI-driven methods, with a comparison of their effectiveness in different software development contexts. Finally, the paper discusses the challenges and limitations of current prioritization models and suggests future directions for improving prioritization techniques in regression testing.

DOI: https://doi.org/10.5281/zenodo.18721641

 

Published by: vikaspatanker

Authors: Ms. Meenakshi, Dr. Shweta Mishra

Abstract: Regression testing plays a critical role in ensuring the continued functionality and quality of software as changes, updates, or bug fixes are implemented. Given the complexity and the growing size of modern software systems, managing large test suites efficiently is becoming increasingly difficult. This paper focuses on the importance of test case prioritization, a technique aimed at ordering test cases in a way that maximizes fault detection early while minimizing resource utilization. The study compares various theoretical models and methods of test case prioritization, including random, requirement-based, code-based, and model-based prioritization. It highlights the benefits of prioritization in improving test efficiency and the challenges associated with its implementation. Additionally, the paper reviews both traditional and advanced prioritization techniques, including risk-based prioritization and AI-driven methods, with a comparison of their effectiveness in different software development contexts. Finally, the paper discusses the challenges and limitations of current prioritization models and suggests future directions for improving prioritization techniques in regression testing.

 

 

 

Published by: vikaspatanker