Authors - Rachit Chetankumar Mehwala, Angshuman Kishore Mahato, Patil Sujit Maruti, Surendra Solanki, Gaurav Kumawat, Ravindra Kumar Soni Abstract - Disaster whether it is natural or man-made most of time led to significant challenges to society, often resulting in loss of life, economic instability, infrastructure damage. Effective "Disaster Management" requires seamless communication and good connectivity under extreme conditions. 6G technology with its capabilities such as ultra-low latency, large bandwidth, and terahertz frequencies offers an evolutionary approach to real-time disaster management. This paper explores how 6G technology can be harnessed to build reliable system capable of alleviating the impacts of disasters. By using 6G’s unparalleled communication capabilities we propose framework that ensure strong connectivity and efficient resource allocation in real-time and monitoring during disasters.
Thursday August 27, 2026 9:30am - 11:30am IST Virtual Room EGOA, India
Authors - Kumkum Saxena, Akshay Rathod, Shagun Gupta, Archie Shah, Deep Prajapati Abstract - Professional scarcity, the absence of individualized treatments, and access to mental health assistance only in limited regions creates problems in mental health care. These issues can fundamentally be solved with the introduction of AI. AI has powerful applications in the realm of mental health care, including systematic classification and analysis of data, as well as facilitating tracking and predictive treatment that leads to personalized medicine. Chatbots, predictive analysis and cognitive computing try to give precise diagnosis by facing the unsolved challenges in the empiric domain of cognitive sciences. This review attempts to highlight the need for multidisciplinary collaboration and more research so that mental health care AI systems are more inclusive.
Authors - Amruta Amuna, Hardik Rokde, Anuj Gosavi, Gaurang Gulhane, Ishaan Chepurwar, Arnav Jadhav, Vinayak Musale Abstract - In today’s digital age, most of the work has become sedentary. Whatever you want to do is now available at the tip of your fingers, reducing your physical activity and causing individuals to suffer from chronic health issues, such as back pain and postural disorders. Yoga has been publicised throughout the world, and now all of us are aware, and many of us even perform yoga regularly. But doing yoga is not enough. We must do it efficiently & with the correct posture to get the benefits of yoga. It has been observed that more than 40% of people doing yoga do it incorrectly. This motivated us to develop a smart, Iot-enabled solution that addresses this problem during yoga practice. The system integrates Force Sensing Resistors (FSRs) to measure the force applied on the sensor, LED indicators to guide the correct position for palms and feet for every yoga pose, and a buzzer to give auditory feedback for misalignment. By combining all of these along with the Arduino Uno microcontroller board, this solution bridges the gap between self-guided yoga practice and expert supervision, making each yoga pose more efficient, easier, and more effective for the user’s body.
Authors - Sahana S, Umabharati H, Rakshita.G, Vaishanvi.K, Nikita Patil Abstract - In a populous nation like India, one fundamental need is travel so travel encompasses road, rail and water. Road transport is most utilized and also the prime reason people get to lead simple lives. Even after studying the present situation, we found a major problem on the roads: potholes. These potholes have become a source of harm to the condition of the roads and an additional threat of accidents on the roads. The detection of potholes is vital for safety on roads. The best method for pothole detection is using the real-time accurate efficient YOLOv10 model. A Raspberry Pi Camera Module can record real-time video and images of the road. Further, the Raspberry Pi can be integrated with a GPS module to find the precise coordinates of any potholes. The data generated by the GPS module is helpful in making repairs and guiding drivers in choosing routes. The system relies on a Convolution Neural Network (CNN) model, which assists in pothole detection using YOLO models.
Authors - Jai Ramani, Tanay Kelkar, Darshil Shah, Divanshu Maheshwari, Archana Nanade Abstract - Anonymized employee reviews on platforms like Ambition-Box offer insights into workplace experiences such as salary, work culture, working hours, and management quality. However, manually analyzing large volumes of reviews is challenging and time-consuming. To overcome this limitation, an automated system is proposed to collect, process, and present employee sentiment in a structured and meaningful way. Using the technique of Aspect-Based Sentiment Analysis (ABSA), the system classifies reviews as positive or negative while identifying sentiment across key concerns such as salary, work-life balance, career growth, management quality, etc. To identify the keywords and their corresponding sentiment, this study utilizes the T5 model that is fine-tuned using the InstructABSA framework. Data is gathered through web scraping, ensuring coverage of employee opinions from multiple platforms. The resulting analysis highlights areas where companies excel or need improvement, providing actionable insights to enhance the workplace.
Authors - Dhaval Shah, Shivani D. Anjaria, Bhupendra Fataniya Abstract - Hardware Security is the key aspect of the integrated circuit’s life cycle; Any malicious modification in the system design at the foundry is a significant concern for hardware threats, known as a Hardware Trojan attack. These Trojans are very difficult to detect in the real world, even during manufacturing and testing. In this article, the impact of Hardware Trojan on the performance of the cache memory is presented. Insertion of Trojan demonstrated in the cache replacement policy, which replaces the original cache replacement (least recently used, first in first out, and least frequently used) policies with another replacement policy (most recently used). The performance was analyzed in the gem5 simulator after inserting a Trojan. It was clearly evident that Trojan insertion degrades cache performance and affects overall processor performance. It is observed that the impact on the Trojan was a savior on LFU compared to other cache replacement policies, since LFU persists in its counter-based memory.
Authors - Ajay Talele, Revati More, Satej Patil, Shreya Bedre, Gokarn Nemade, Harshwardhan Vanmore, Dipak Parvate, Sakshi Dhumale, Varun Deshmane, Vedika Dange Abstract - An Advancement in Effective Parking Solutions: The Smart Car Parking system. In cities, parking congestion results in wasted time, fuel, and irritated drivers. By offering real-time parking availability updates, expediting the procedure, and lowering traffic in parking lots, the Smart Car Parking System provides an answer. The system, which was constructed with an Arduino Uno microprocessor, uses infrared (IR) sensors to identify whether a car is in each slot. To assist drivers in making educated judgments, a linked LCD shows real-time data on available and occupied spaces. Entry and exit barriers are controlled by servo motors, which grant only permitted access. The technology automatically updates the slot status when cars enter or exit. This improves traffic flow, lowers pollutants, saves fuel, and lessens the need for manual supervision. The system is perfect for public lots, workplaces, malls, and residential areas. It can be improved with AI for space optimization and IoT-based apps for slot reservations, increasing the sustainability and efficiency of urban parking. [5]
Authors - P Sanjana, Smrthi Harits, Divyadarshan C.S Abstract - This project aims to solve the issues faced by Bharatanatyam dancers in accessing the translations and interpretations of compositions used for Bharatanatyam performances. Understanding these compositions is crucial in depicting an accurate vision of the composition. Through structured user interviews, it was discovered that due to the non-preservation of these translations and limited access to scholars, accurate translations, and music resources, dancers face significant issues. This hinders their creative process and restricts their creative freedom. Compared to experienced dancers with access to vast resources, the upcoming artists have very few such connections and guidance, making them more vulnerable to this problem. This necessitates a solution to reduce the accessibility issue faced by the upcoming dancers. Various design methodologies were employed to tackle the issues faced by Bharatanatyam dancers. To bridge this gap, Kalaahithaa, a digital service platform, was designed to provide a repository of verified translations, connect dancers to scholars, musicians and teachers to help them understand the compositions, create new compositions and choreographies, and foster collaborative opportunities. This solution was further developed using a service design blueprint, user flows, site maps, and low-fidelity and high-fidelity prototypes. The platform supports personalised interactions, enabling dancers to seek expert guidance, upload and access translations, and engage in meaningful exchanges that enhance their understanding of compositions while promoting monetary ethical practices. By centralising resources and fostering connections with a usercentred approach, Kalaahithaa serves as a vital tool for dancers to refine their art while preserving the integrity of Bharatanatyam.
Authors - Sonali Antad, Kalyani Ghuge, Prakash Sharma, Vaishnavi Chirawande, Aditi Gade, Shweta Ahire, Sharvari Jadhav Abstract - This study describes an AI healthcare chatbot that automatically assesses patients’ medical needs, conducts inter- views and performs thorough health analyses. Using several state-of-the-art natural language processing (NLP) models, including sentence transformers and Llama-based large language models, the system analyzes user symptoms and classifies them into specific medical domains like diabetes, blood pressure, skin and stomach disorders. It uses Gemini API’s generative AI. This dynamically creates several questions and suggests some medical diagnoses. The platform’s improved fishbone diagram considerably aids root cause analysis by visually showing how potential causes relate to user-reported symptoms. Our project creates efficient, interactive healthcare assistants that improve access to preliminary medical advice.
Authors - Rhucha Deodhar, Tanya Gadwal, Ananya Bhat, Aditi Hinge, Shilpa Pant Abstract - This paper presents a real-time, vision-based system for Indian Sign Language (ISL) recognition and translation, aimed at enhancing communication between the deaf community and non-signers. The system combines a CNN-LSTM architecture for static gesture recognition, achieving an accuracy of 98.47% and introduces GestureNet, a bidirectional LSTM model trained on a custom dynamic gesture dataset, which attains 96.83% recognition accuracy. Ad-ditionally, a Generative AI framework is integrated to convert recognized ges-tures into semantically coherent and contextually appropriate sentences. By em-phasizing real-world applicability and high recognition performance, the pro-posed system advances sustainable and accessible communication technologies, with potential impact in education, public services, and digital inclusion, partic-ularly in developing regions.
Authors - Arokiaraj S, Amudha T, Swamynathan R Abstract - Technology has become the driving force of progress and development all around the world. The recent development of Generative Artificial Intelligence has led to a revolution in the field of Education, Employment and Human Resources. Securing a dream job or building a suitable career is the goal of every student. Likewise finding the right candidate for the job is the goal of every employer. LLMs (Large Language Model) comes to the rescue, through dynamic question content creation for a selected topic and test the candidate for that set of skills. A candidate’s unique set of skills and abilities are understood and tested by the Generative AI where conventional methods fall short to meet this criterion. This paper proposes an automated question generation framework built using LangChain, LLM model -GPT Turbo 3.5 from OpenAI API and Streamlit application development tool. This application successfully tests the skills of the candidates by asking customized multiple-choice, true/false and open-ended questions based on their chosen topic, knowledge level, number of questions and time limit. Results indicate that this framework can create a challenging environment for the contenders thereby facilitating the interview process and selection of highly suitable candidates.
Thursday August 27, 2026 12:30pm - 2:30pm IST Virtual Room EGOA, India
Authors - Shailaja Uke, Mohit Garg, Suyash Chandolikar, Swayam Chandak, Shriraj Nelekar Abstract - Chest X-rays are the most common tool for diagnosing various thoracic diseases. However, manual interpretation is time-consuming and prone to human error. This paper presents a deep learning approach for automated pathology detection in CXRs using the Customized DenseNet-121 model. The model performs binary classification to identify 14 pathologies, including cardiomegaly, pneumothorax, mass, and edema. To address class imbalance in medical imaging datasets, weight normalization is applied. Additionally, the visualization technique of Grad-CAM enhances interpretability by pointing out the most critical regions influencing the model's decisions, which helps healthcare practitioners assess. Toward further refinement of segmentation and improvement in precision of localization, we incorporate a customized U-Net model to enhance better delineation of regions of interest. Our model achieves an overall AUC of 87%, showing the highest accuracy. The customized U-Net integration improves seg-mentation performance, reducing localization error by 15%. This approach not only enhances diagnostic accuracy but also provides transparent decision-making, making it a valuable tool for medical professionals.
Authors - Aditya Gaura, Mandeep Kaur, Kimmi Verma, Monali Gulhane, Nitin Rakesh Abstract - This cutting-edge research paper introduces a paradigm shift in parking management, underpinned by an intricate network of technology and user-centric design. The system's hallmark feature is its advanced slot allocation mechanism. Users can make reservations via the mobile or web application, with the system autonomously assigning slots based on a holistic evaluation of user characteristics, such as vehicle type and duration of stay. Leveraging IoT integration, the system employs a sophisticated array of sensors and cameras to monitor parking slot occupancy in real-time, resulting in a fluid entry and exit process. The user experience is paramount in this system. It offers a tailored approach based on user type, streamlining the process for faculty, students, and visitors. Notably, the reduction in the time spent hunting for parking spots has the potential to mitigate the perennial issue of urban traffic congestion. This, in turn, aligns with environmental conservation efforts, as the system indirectly lowers emissions and the carbon footprint associated with circling for parking spaces. Moreover, this system's role as a data aggregator is invaluable. It collects and processes a wealth of data, offering parking operators unprecedented insights into daily usage patterns, peak periods, and favored slots. This data-driven approach empowers operators to make informed decisions about slot management, maintenance, and resource allocation.
Authors - Balwinder Kaur, Jaswinder Singh, Deepika Abstract - Nature-inspired optimization algorithms constitute a class of computational techniques that derive their underlying mechanisms from biological, ecological, and physical systems. By emulating processes such as evolutionary adaptation, collective swarm behavior, and decentralized decision-making, these algorithms offer robust solutions to complex optimization challenges across engineering and computational domains. Notable methodologies include Genetic Algorithms, Particle Swarm Optimization, and Ant Colony Optimization, each demonstrating efficacy in handling both single and multi-objective optimization problems, including those involving high-dimensional search spaces and non-linear constraints. Within the field of Automatic Speech Recognition (ASR), nature-inspired optimization techniques are instrumental in refining critical system components. Their application spans feature selection, acoustic model training, language model optimization, and efficient decoding strategies. By leveraging adaptive search mechanisms, these algorithms enhance model accuracy, reduce computational overhead, and improve generalization in ASR systems. This research study presents a systematic examination of nature-inspired optimization methods, focusing on their theoretical foundations and practical implementations in ASR. Furthermore, it critically evaluates existing challenges, such as sensitivity to hyperparameter tuning, computational scalability with large-scale datasets, and the absence of comprehensive convergence guarantees. Addressing these limitations is essential for advancing the applicability of nature-inspired optimization in next-generation speech recognition systems and related domains.
Authors - Anchal Saini, Nitin Kulshrestha Abstract - In the rapidly digitizing financial landscape, the ability to effectively use digital tools has become essential for financial well-being. This study examines the impact of digital competence on financial resilience within dual-income married couples, adopting a dyadic perspective. Drawing on the Actor-Partner Interdependence Moderation Model(APIMoM), it studies both actor and partner effects of digital competence. As well as the moderating role of each partner’s attitude toward FinTech. Data was collected from 107(214 individuals) working couples in Gurgaon, India. Covariance-Based Structural Equation Modeling using SmartPLS revealed that digital competence significantly influences both individuals' and their partners’ financial resilience. Moreover, attitudes toward FinTech were found to moderate these relationships, strengthening the positive effects of digital competence. Notably, the husband’s attitude had a stronger moderating impact on the wife’s resilience than vice versa, indicating potential gender-based dynamics. The study marks the importance of addressing both digital skills and relational attributes in aiding household financial resilience. Practical implications suggest that digital literacy programs should consider couple-based interventions that target both digital competence and attitude change.
Thursday August 27, 2026 12:30pm - 2:30pm IST Virtual Room EGOA, India
Authors - Rasika Ransing, Kaushik Sakre, Neha Kudu, Shivam Shinde, Siddhi Talkar Abstract - The introduction of Automated Essay Scoring systems brought better assessment methods into education through standardized scoring systems that operate at scale while being time efficient. The current AES models function exclusively with English content while neglecting multilingual evaluation, particularly in the Hindi and Marathi languages. A multilingual AES framework has been developed using transformer models XLM-RoBERTa, MuRIL, DistilBERT, and mBERT for conducting context-based essay assessments throughout English, Hindi, and Marathi texts. Through multilingual embeddings combined with fine-tuned models, the system maintains cohesive and coherent, and argumentative quality in essays. The assessment by QWK and RMSE metrics demonstrates both high accuracy and reliability of the system. The highest performance emerged from XLM-RoBERTa and Google MuRIL at 0.78 QWK and 0.77 QWK, respectively.
Authors - Sarvesh Shinde, Tarun Kurakula, Venugopal Murugan, Tanmay Patil, Aparna Bannore Abstract - Network security is a difficult topic these days, with threats appearing quickly and everywhere. According to the study "Network Security Using Graph Embedding," connections are visible when jumbled network data is transformed into transparent graphs. First-order graphs have direct node links, while second-order graphs have nodes that share neighbours. DeepWalk, Node2Vec, and sense-making tools. Node2Vec, choice-based, tight groups or large network view, and modified random walks. DeepWalk is a straightforward, sequential structure mapping method. Both embeddings are feasible in terms of network layout. A graph as opposed to the outdated equal-link techniques, Attention Network, GAT, and anomaly hunt use attention tricks for important connections. fresh activity in the dataset, labeled data, normal versus odd markers, and real-time data. Strange spikes, unusual nodes, rules established, and threats identified. In continuous networks, not data crunch for kicks, quick catch, hackers, or weak spots.
Authors - Atharva Shirbhate, Jay Sutar, Om Tathed, Bhupal Shelke, Geeta S. Navale Abstract - The development of Artificial Intelligence (AI) has led to significant advancements across numerous domains, including finance, healthcare, and customer service. Recent progress, particularly in the field of Natural Language Processing (NLP), has been driven by the emergence of Large Language Models (LLMs). These models utilize transformer architectures and vast datasets to perform a wide range of tasks, such as language translation, text generation, and complex data analysis. As AI technology continues to evolve, it offers the potential to streamline decision-making processes, enhance data management, and provide personalized recommendations. This study focuses on leveraging AI to address specific challenges in the financial sector. The objective of this paper is twofold: firstly, to develop a system capable of recommending bonds based on user-specific requirements, thus aiding investors in making informed decisions; and secondly, to collect bond data from sellers and integrate it seamlessly into an existing database. Through a systematic review of recent AI advancements and prompt engineering techniques, this paper aims to provide insights into how these technologies can be harnessed to improve financial data integration and recommendation systems
Authors - Kumkum Saxena, Ayesha Nagdawala, Esha Nemani, Jatin Mawa, Mamta Gupta Abstract - Even in the modern days of digital era, reporting crimes like those of corruption or misconduct is still difficult owing to the fear of retaliation. Some traditional reporting mechanisms are available, but they often do not allow enough anonymity or security, preventing tipsters from reporting. Many whistleblowers face serious consequences, including job loss, legal action, or even physical threats, making them reluctant to report wrongdoing. SHADE is a blockchain-based solution that seeks to overcome these challenges by providing a decentralized and tamper-resistant medium for anonymous tip-offs. SHADE stands apart from conventional systems, which store centralized databases vulnerable to breach, providing full anonymity and data integrity with encryption. This paper explores SHADE’s architecture, which integrates blockchain for immutable data storage, cryptographic encryption for secure communication, and smart contracts for automated processing.
Thursday August 27, 2026 12:30pm - 2:30pm IST Virtual Room EGOA, India
Authors - Suhas Bhise, Ketki Kshirsagar, Vivek Deshpande Abstract - In the context of Industrial Edge Computing, the growing deployment of IoT and mobile devices has resulted in an explosion of real-time, high-velocity time series data. This paper investigates statistical and machine learning approaches for time series forecasting in such environments, where latency, bandwidth, and computational efficiency are critical constraints. We evaluate traditional methods like Simple Moving Average (SMA), Holt-Winters Exponential Smoothing, and ARIMA, and contrast them with machine learning models such as Logistic Regression and XGBoost. Experiments conducted on the Microsoft Azure Predictive Maintenance dataset demonstrate that SMA and ARIMA offer comparable baseline accuracy, while XGBoost outperforms them in terms of forecast quality for multivariate series. We also explore the effectiveness of SMOTE for improving failure prediction using logistic regression. The findings suggest that lightweight models like XGBoost with lag feature engineering can be viable for forecasting in edge environments.
Authors - Vedant Chandore, Niranjan Pardeshi, Sai Sinare, Samruddhi Akude, Sahil Dhawane, Kartik Gawande, Rahul Sadgir, Shravani Nigade, Ajay Talele Abstract - For transportation infrastructure to be safe, effective, and long-lasting, road condition monitoring is essential. Conventional techniques, which depend on human inspections, are frequently ineffective and prone to mistakes. To overcome these constraints, this study suggests a machine learning-based smart road condition monitoring system. Utilizing cameras installed on vehicles, the system gathers pictures and videos of the state of the roads, which are subsequently processed by sophisticated machine learning algorithms. These algorithms categorize surface conditions, identify irregularities in the road, and offer information on repair requirements. Through comprehensive field testing and data analysis, the study shows how effective the system is, showing notable gains in both the efficiency of maintenance procedures and the accuracy of identifying road issues. By concentrating on image and video analysis, this smart monitoring system offers a revolutionary solution for urban infrastructure management, opening the door for safer, more intelligent, and sustainable road maintenance procedures. This strategy not only lowers operating costs but also improves road safety and infrastructure sustainability.
Authors - Sai Himagnya Parisaneni, Vemula Surya Teja, Revanth Guthula, Sushama Rani Dutta Abstract - This study presents an optimized approach for detecting mental disorders by integrating support vector machines (SVM) enhanced through Minimum Bayes Error Rate (MBER) optimization. The proposed framework uses MBER Optimization and refines classification boundaries through SVMs improve decision-making. Unlike conventional deep learning approaches that rely solely on CNN based end-to-end learning, our method uses SVM for classification that minimizes errors, enhancing model robustness and generalization. The experimental evaluation on EEG-based datasets assesses the effectiveness of the hybrid approach in terms of accuracy, computational efficiency, and scalability. The results provide insights into the potential of MBER-optimized SVM models for real-world applications in mental health diagnostics.
Authors - Satish Chikkamath, Shreya Pattanashetti, Pooja V Gadad, Vidya Revanakar, Bhoomika Hosamani Abstract - Emojis serve as an established means for people to express emotions and sentiments while interacting on social media. This paper examines the task of emoji prediction from text by developing accurate classification methods. The model uses a pre-trained and fine-tuned BERT framework on a dataset consisting of text sentences along with their corresponding emojis. This structured data allows the model to capture contextual meaning and emotional nuances, which are crucial for practical applications. Challenges associated with emoji usage are addressed through tokenization techniques in text preprocessing, while performance advances are achieved using stemming and feature extraction. Research conclusions indicate that the BERT-based model outperforms traditional deep learning approaches like LSTM. This study spotlights how NLP and sentiment analysis contribute to emoji prediction and shows its practical applications in social media monitoring, sentiment analysis, and enhancing user experiences.
Authors - Sagar Janokar, Krish Deshpande, Krishna Masane, Shriyash Kothe, Varad Kulat, Krish Chabria, Rushikesh Kuchekar Abstract - This project demonstrates how a machine learning based approach can revolutionize the analysis of unstructured text data in defense intelligence. By automating key processes, the system will enable faster and more accurate identification of threats and patterns, improving decision-making and operational efficiency. This innovative application highlights the transformative role of technology in addressing real-world challenges in intelligence gathering.
Authors - Dhanaselvam J, Dhanalakshmi R, Prashaanth S, Hariprasath S, Harish R Abstract - India, with three-fourths of its population dependent on agriculture, is plagued by severe crop loss due to pest infestation, particularly in staple crops like rice, wheat, maize, and soybeans. This paper proposes an embedded system of real-time pest detection and precise pesticide spraying to enhance productivity. The system employs deep learning with a Residual Neural Network (ResNet) and Quadra-attention, residual, and dense fusion techniques for enhanced pest image classification. High-resolution images of crop leaves are captured, pre-processed, and analyzed for pest detection. Upon detection, the system selects the appropriate pesticide and activates an autonomous robotic sprayer. Driven by an Arduino NANO-based module with an L293D motor driver, the robotic system automatically navigates through fields, ensuring precise pesticide application without waste and infrastructure costs. With IoT integration and 99.80% validating accuracy, this system optimizes pesticide use, enhances crop health, and enhances yield, offering a cost-effective automated pest management system for sustainable agriculture.
Authors - Nikita Bhatt, Nirav Bhatt, Purvi Prajapati Abstract - In today’s data-rich world, we often deal with multiple types of information such as images, text, and audio. Traditional deep learning models usually focus on a single type of data, but real-world applications need systems that can understand and connect across these different formats — a concept known as multi-modal learning. This paper explores cross-modal retrieval, where a user can input one type of data (like an image) and retrieve another (like related text). To make this possible, we map different data types into a common space using deep learning methods like CNN for images and LSTM for text. One of the key challenges in this area is comparing vectors of different lengths, which affects similarity estimation. Most traditional methods use inner product similarity, which is not ideal for vectors with varying magnitudes. To overcome this, we normalize the vectors using cosine similarity, which focuses only on the angle between vectors, not their length. This improves retrieval accuracy by reducing noise caused by vector size differences. We also discuss the benefits of using deep learning to jointly learn features and generate hash codes for faster and more accurate retrieval. Experiments on datasets like Google News show that cosine similarity outperforms Euclidean distance in terms of retrieval performance, especially when combined with models like CBOW.
Authors - Ria Ashish Gawali, Christopher Sachin Chopde, Aryan Gupta, Tashmeet Kaur Jasbeersingh Hora, Rachna Karnavat Abstract - MediaGPT is a Generative AI system that combines natural language and image synthesis to create unified, visually appealing media content. By combining strong language models such as ChatGPT and Phi-3 with image synthesis models such as Stable Diffusion and ControlNet, MediaGPT facilitates intelligent text-image alignment on an interactive canvas. The layout can be easily customized along with semantic coherence and aesthetic balance. Developed for designers, educators, marketers, and content creators, MediaGPT improves the creative process by facilitating effortless multi-modal integration and providing easy-to-use tools for creating high-quality, contextually appropriate content.
Authors - Lokesh Khedekar, Atharva Kassa, Kartavya Sharma,Tejas Kedar, Sarthak Kasar, Kaustubh Kelgandre, Sharad Kasralikar Abstract - Natural Disasters have been a major threat to the living beings, environment and the infrastructure, in mainly areas where they have poor access to early warnings systems. This paper provides AI-based Natural Disaster Response System which helps to evaluate the impact of natural disasters and gives better of the existing systems. The system has historical Geographic Information System (GIS) datasets with real-time data from Internet of Things (IoT) sensors and predictive modeling to check out the natural disaster’s magnitude, area of impact, and resources. The methodology includes data preprocessing, feature extraction, and machine learning model training to achieve effective predictive accuracy. A Convolutional Neural Model (CNN) model was created and tested which further achieved 93% accuracy of predicting the impact of the disaster incident. The system was then compared with other machine learning models, then was proved to be more effective. The suggested method gives efficient, cost-effective and scalable way of utilizing the emergency resources at the maximum.
Thursday August 27, 2026 3:30pm - 5:30pm IST Virtual Room EGOA, India
Authors - Prasanna Lakshmi T, Shankar Lingam. M Abstract - This paper explores the intersection of emerging technologies and ICT policy evolution in India, with a focus on Artificial Intelligence (AI), blockchain, the Internet of Things (IoT), and 5G technologies. As India navigates its digital transformation through initiatives like Digital India, the paper examines how the nation's ICT policy framework has adapted to accommodate these disruptive technologies. Using a theoretical approach based on Technological Innovation Systems (TIS), the study traces the historical development of India's ICT policies, from early telecom regulations to the modern-day focus on digital infrastructure and smart technologies. Challenges such as the digital divide, cybersecurity, and data privacy are also analyzed. By identifying key policy milestones and evaluating India's current efforts in integrating emerging technologies, this paper provides insights into the future direction of ICT policy in India. The findings highlight both opportunities and barriers to sustainable technological advancement and offer policy recommendations to better align ICT governance with global trends.
Authors - Agatsya Yadav, Renta Chintala Bhargavi Abstract - Large Language Models (LLMs) offer powerful capabilities but their significant size and computational requirements hinder deployment on resource-constrained mobile devices.This paper investigates Post-Training Quantization (PTQ) for compressing LLMs for mobile execution. We specifically apply 4-bit PTQ using the BitsAndBytes library via the Hugging Face Transformers framework to Meta’s Llama 3.2 3B model. The quantized model is further converted to the GGUF format using llama.cpp tools for optimized mobile inference. The proposed PTQ workflow achieved a 68.66% reduction in model size through 4-bit posttraining quantization, enabling the Llama 3.2 3B model to run efficiently on a standard Android device. Qualitative validation confirmed the 4- bit quantized model’s ability to perform inference tasks successfully. We demonstrate the feasibility of running the final quantized GGUF model on an Android device using the Termux environment and the Ollama framework. PTQ, particularly down to 4-bit precision combined with mobile-optimized formats like GGUF, presents a viable pathway for deploying capable LLMs directly on mobile devices, balancing model size and functional performance.
Thursday August 27, 2026 3:30pm - 5:30pm IST Virtual Room EGOA, India
Authors - Dwayne Nixon, Shaun Menezes, Ramya Kulkarni, Phiroj Shaikh Abstract - In today’s fast-paced development environment, where efficiency and speed are paramount, manual tasks such as taking screenshots, converting files, rebooting systems, and managing repositories have become increasingly tedious and time-consuming. These routine activities disrupt developer workflow and hinder productivity, consuming valuable time. To address these inefficiencies, this work proposes a comprehensive automation tool that extends beyond handling basic tasks to streamline workflows and optimize productivity. Firstly, this tool centralizes a wide range of operations, including automating code generation, creating detailed reports, and developing websites. By integrating these functionalities, developers can eliminate redundant tasks and focus on high-level problem-solving. Secondly, the automation tool enhances accuracy and consistency across development projects, ensuring higher standards of work and reducing errors associated with manual processes. Furthermore, the tool aligns with evolving technological demands, enabling teams to adapt to increasing project complexities while maintaining efficient workflows. This solution represents a transformative approach to software development, combining automation and centralization to reduce manual workloads and optimize developer productivity. The implementation of such an all-in-one automation platform promises to significantly improve efficiency and foster innovation in the industry.
