Authors - Deepti Maheshwari, Shailesh Gahane Abstract - Website phishing poses a massive security threat that continues to increase in prevalence. Internet scammers exploit human faith by running imitation websites which aim to obtain confidential user information. An extensive review of multiple phishing detection techniques and hybrid detection models appears in this paper which brings together different detection methods to speed up and increase the accuracy of breaking down phishing-related websites. The paper investigates how machine learning (ML), artificial intelligence (AI) and heuristic-based approaches and anomaly detection should be implemented within hybrid systems which detect phishing behavior. Multiple studies from the literature receive analysis through which we identify their research approaches as well as their outcomes together with their limitations along with their contributions to the field. The evaluation will demonstrate how hybrid models can boost the detection of phished emails while detailing methods to strengthen model performance and flexible design and growing capability.
Authors - Rinkle Solanki, Shailesh Gahane Abstract - Surgical site infections (SSIs) represent a major concern for the healthcare industry, highlighting the need for timely intervention and effective predictive strategies.This paper presents an external validation framework for machine learning algorithms designed to identify and forecast SSIs in advance. We use sophisticated algorithms to create accurate predictive models using a variety of variables, including clinical features, microbiological data, and patient demographics. Across a range of patient demographics and therapeutic circumstances, thorough external validation is carried out. Our results demonstrate how effective this strategy is at precisely identifying SSIs, enabling prompt interventions, and improving patient outcomes. Surgical care procedures could be improved and medical expenses could be decreased by using validated models.
Authors - G Jeyashaathvee, Anish Pranav, Sindhu Chandra Sekharan, Jesline D, Ajanthaa Lakkshmanan Abstract - The challenge of extracting useful insights from unstructured data in the presence of big digital information is still present today. AI Intensive Timestamp based Summarization proposes a novel framework to automatically extract and summarize pivotal events, along with their corresponding timestamps from different data sets. Making use of Natural Language Processing and deep learning machine learning approaches, the system checks text data for temporal markers, extracts salient events and produces verbose summaries. The method proposed shall make the historical analysis easier and trend detection along with automated reporting of large dataset summaries more concise. Technique is implemented using Named Entity Recognition for date extraction and transformer models in case of summarization Experiments show that AI-based timestamp summarization is efficient in enhancing IR results and autodoc reliability. Making This Research Unique and Contributing to the much Expanding AI-driven text analysis filed, a Scalable in nature for timestamp extraction on domain wise basis.
Authors - Ankita Mehta, Shailesh Gahane Abstract - This paper discusses about, we display a profound learning-based approach bipolar clutter discovery utilizing Convolutional- Neural Systems (CNN) and Long Short-Term Memory (LSTM) systems. To assess the model’s performance, two particular datasets Twitter information and survey data were analyzed. Preprocessing steps, counting information enlargement, normalization, and the application of the Adam optimizer, were joined to upgrade the model’s adequacy. The model’s exactness and misfortune were measured for both datasets, and it was watched that the survey dataset given superior execution, yielding higher precision and lower misfortune compared to the Twitter dataset. These discoveries propose that the questionnaire-based information may be more reasonable for solid bipolar disorder location within the given show. The inquire about illustrates the potential of combining CNN and LSTM for mental wellbeing examination, highlighting the significance of information determination in accomplishing ideal comes about.
Thursday August 27, 2026 9:30am - 11:30am IST Virtual Room BGOA, India
Authors - Sarika G. Songire, Deepa S. Deshpande Abstract - Deep learning algorithms have completely transformed medical diagnostics by enabling accurate and efficient identification of brain tumors. Brain disorders often arise due to increased in excessive and improper cell counts, which can damage neural structures and, in severe cases, lead to malignant brain cancer. The reducing mortality rates requires prompt intervention and early discovery. This research work presents an architecture of deep neural network specifically designed for the purpose of detecting brain tumors from MRI images. The proposed model is evaluated against existing research using a similar dataset to assess its effectiveness. For comparison, performance parameters including area under the curve (AUC), recall, accuracy, precision, and loss are used. According to experimental results, the suggested CNN model accomplishes 98.61% accuracy, 99.70% AUC, 99% of both precision and recall, and 0.41 loss in a data set of 3,264 MRI images. These findings indicate that the suggested model surpasses current models and provides a dependable and effective technique for prompt brain tumor diagnosis.
Authors - Ayush Itkhede, Aryan Fulsunge, Sarthak Bomanwar, Sujal khobragade, D.M.Shinde Abstract - The integration of embedded systems into brick manufacturing has led to significant improvements in efficiency, quality control, and resource management. This paper presents an embedded system designed to optimize brick production by monitoring critical parameters such as temperature, humidity, and material composition in real-time. The system achieved a 32% reduction in production cycle time, from 6.5 hours to 4.4 hours per batch, and improved material handling speed by 41%, from 120 kg/hour to 169 kg/hour. Quality control metrics showed a 78% reduction in defect rates, from 5.2% to 1.15%, while energy consumption decreased by 29.3%, from 17.4 kWh to 12.3 kWh per 1,000 bricks. The system also reduced waste by 64%, from 8.9% to 3.2%, and improved material utilization rates from 83.5% to 94.8%. These results demonstrate the potential of embedded systems to revolutionize brick manufacturing, making it more sustainable and cost-effective.
Authors - Kamini Solanki, Rahul Vaghela, Jay Panchal, Anjali Mahavar, Jaimin Undavia, Nilay Vaidya Abstract - Diabetes is a chronic illness that affects the body's ability to metabolize glucose, which is the sugar that provides energy to the body. Type 1 and type 2 diabetes are the two main types of the disease. The immune system of a person with type 1 diabetes attacks and destroys the pancreatic cells that make insulin, which causes blood sugar levels to rise. Type 1 diabetes symptoms include excessive thirst, frequent urination, extreme hunger, weight loss, fatigue, impaired vision, delayed healing, tingling or numbness in the hands and feet, and recurring infections. AI algorithms may be used to collect and analyze medical data to support early diagnosis and treatment. Typically developing in childhood or adolescence, type 1 diabetes can be managed with the injection of insulin or an insulin pump. A person with type 2 diabetes either develops an inability to use insulin or ceases making enough of it to regulate blood sugar levels.
Authors - Dipali Wankhade, Shailesh Gahane, Mrunal Meshram Abstract - Oral cancer is one of the fatal diseases present in society. Its late-stage diagnosis puts it under the list of diseases with high mortality rates. Deep learning has conceptualized the revolution in the field of medical imaging with impressive depth and accuracy in diagnosis and early detection. This review highlights the advancements in deep-learning-based methodologies for oral cancer detection and classification, including convolutional neural networks (CNNs), recurrent neural networks (RNNs), deep reinforcement learning (DRL). Multi-modal learning and hybrid models combining histopathological and radiological data have increased the precision of tumor segmentation and subtype classification. The observational data will help provide insights into the factors responsible for explaining the model's decisions and the associated risks, which are most relevant for potential applications. The development of transfer learning and self-supervised learning also seems to have significantly solved some of the most serious challenges regarding the volume of clinical data. Future studies can harness standardized practices for data collection, deploy technically sound explainable AI frameworks, and conduct clinical validations as the closing gap between tremendous strides made in deep learning and practical real-world implementation. This review provides a wide-ranging overview of such AI-driven methodologies, focused on the critical challenges and future directions for improvements in early oral cancer detection and reduced mortality rates.
Authors - Ankita Mehta, Shailesh Gahane Abstract - In this research, we focus specifically on mental disorder which is bipolar disorder using machine learning techniques, utilizing a simple dataset from Kaggle for training and evaluation. The study involves applying various ML models, including R. Forest, XG-Boost, and S. Vector Machines (SVM), on the dataset. We assess the performance of these models using evaluation MSE (how far actual value to predicted value), Precision, and Fl Score to determine their effectiveness in predicting bipolar disorder. However, through extensive experimentation, we found that the combination of (CNN) and (LSTM) networks outperformed the other algorithms, achieving an overall accuracy of 95%.
Thursday August 27, 2026 9:30am - 11:30am IST Virtual Room BGOA, India
Authors - Vijeta V Shettar, SR Nirmala, Satish Chikkamath, Suneeta V Budihal Abstract - This study investigates the performance of various speaker recognition models, including SpeakerNet, TiTANet Small, and TiTANet Large, using the IITG-MV dataset for speaker verification tasks. The preprocessing steps involved resampling, noise reduction, segmentation, and volume normalization to prepare the audio data for input into the models. The models were evaluated based on their ability to correctly verify whether two audio samples belong to the same speaker or not. The evaluation metrics, derived from the confusion matrix, revealed that SpeakerNet outperformed both TiTANet variants, achieving an accuracy of 85%, while TiTANet Small and TiTANet Large achieved accuracies of 75% and 80%, respectively. Despite lacking graphical visualizations, the confusion matrix provided a comprehensive view of the models’ performance, showing how each model handled correct and incorrect speaker match predictions. The results highlight that SpeakerNet is the most effective model for speaker recognition in this setup, demonstrating superior accuracy and robustness in identifying speaker-specific features. These findings can guide future research in optimizing speaker recognition models for real-world applications involving speaker verification.
Authors - Madan Kumar Sharma, Nadir Kamal Salih Idries, Abdullah Said Alkalbani, Satyanarayana Degala, Gopal Rathinam, Ankit Sharma Abstract - Multi-resonance (MR) Microwave sensors have emerged as a promising sensing device for mineral-based materials characterization due to their high sensitivity and precision. This research presents a novel microwave sensor for mineral-based material characterization. The sensor's resonating structure comprises a 3 × 3 array of circular-shaped complementary split-ring resonators (CSRR) coupled with four rectangular defected structures etched around the CSRR array. This unique configuration enhances electromagnetic interaction with the material under test (MUT), leading to precise characterization based on S-parameter analysis. To validate the sensor’s efficacy, simulation-based investigations were conducted on the mining-based materials, including chrome, copper, and quartz. The obtained results demonstrate distinct resonance shifts and attenuation variations corresponding to each mineral, highlighting the sensor’s capability to differentiate and analyze their dielectric properties. The proposed MR-sensor design provides a robust and efficient method for non-destructive material characterization, offering potential applications in the mining industry, quality control, and geophysical exploration.
Authors - Geethu Lakshmi G, P. Nagaraj, P. Chinnasamy Abstract - Lung cancer detection constitutes a paramount process in the diagnosis and management of one of the predominant contributors of cancer-induced humanity worldwide. The significance of early screening is underscored by its essential role in enhancing survival rates through the identification of disease during a stage amenable to treatment. Diagnostic methodologies, including imaging modalities, are routinely utilized for diagnosis. Furthermore, advancements in the realm of molecular biology have facilitated the emergence of biomarkers and genetic examines, thereby enabling a more accurate identification of lung cancer. The prompt and defined detection of lung cancer facilitates timely therapeutic interventions, which significantly influence both the efficacy of treatment and overall outcomes for patients. The primary objective of this research is to develop an optimization-based hybrid deep learning methodology for the detection of lung cancer. The initial phase involves pre-processing of input images through techniques such as color space transformation, data augmentation, resizing, and normalization. Subsequently, features derived from Slime Mould Algorithm-based Convolutional Neural Network (SMA-CNN) are employed for the detection of lung cancer, with CNN being trained utilizing SMA, extracted from pre-processed images. Finally, the Squeeze-Inception V3 model, which integrates SqueezeNet and Inception V3, leverages SMA to train the classifier. Consequently, the proposed SMA-based hybrid SqueezeNet-Inception V3 is utilized to classify instances as normal or abnormal. Empirical results designate that SMA-based hybrid SqueezeNet-Inception V3 attained an accuracy of 97.3%, a specificity of 96.1%, and a sensitivity of 98%, thereby underscoring its efficacy in the detection of lung cancer.
Authors - Narayan Gupta, Pawan Kumar, Prince Kumar Singh, Priyabart Kumar, Parampreet Kaur Abstract - For investors, accurately predicting stock market prices is a critical part of financial analysis. This work studies a wide variety of machine learning algorithms specifically designed for the task of predicting stock prices using a variety of techniques and the latest technologies available as of the time of study. The study critically compares a suite of algorithms, including Support Vector Regression (SVR), Random Forests, Decision Tree models, and Long Short-Term Memory (LSTM), each with differing strengths. Moreover, it investigates a variety of approaches that are focused on understanding the complex connections found in the past price data. The dataset used for this study includes extremely long-term stock price representatives over a time range from 2010 to 2024 for Tata Consultancy services (TCS), containing a wide range of information, including opening and closing trading prices, trading volumes, and a variety of calculated indicators reflecting market behaviour. To understand the first experiment carried out for this study, it can be seen that in combination mode, for the best-performing model, Random Forest has the best interpretability. In addition, both Support Vector Regression (SVR) and Decision Tree algorithms deliver both impressive short-term prediction results and clear explanations behind their decisions.
Authors - Sandeep Shinde, Samarveer Moray, Aditya Sakhare, Prathamesh Salokhe, Kedar Sathe Abstract - The exponential growth of digital documents, particularly PDFs, presents significant challenges in efficient information retrieval and extraction. Traditional methods often struggle with the complexity and variability inherent in large PDF documents. Recent advancements in Natural Language Processing (NLP), especially Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG), offer promising solutions. This paper presents a comprehensive system for efficient information extraction from large PDFs using RAG and LLMs. We propose a robust and scalable pipeline addressing challenges such as document segmentation, dynamic retrieval, and response contextualization. Through extensive experiments across multiple domains—including legal analysis, technical documentation, and scientific literature—we demonstrate that the proposed methodology significantly outperforms existing approaches in terms of accuracy, scalability, and efficiency. Our research lays the groundwork for integrating RAG and LLMs in various domains, offering a valuable tool for extracting knowledge from complex documents.
Authors - Rahul Pethe, Parag Puranik, Abhay Kasetwar Abstract - Wireless Sensor Networks (WSNs) have been designed and developed by numerous researchers over time, evolving based on emerging demands and environmental conditions. Over the years, various enhancements and improvements have been proposed, with multiple protocols introduced to address design challenges. In parallel, Mobile Ad Hoc Networks (MANETs) have witnessed tremendous growth in this wireless era. While many researchers have focused on protocols like DSR and hybrid approaches for energy-efficient clustering, these solutions have often proven to be time-bound and context-specific. To overcome these limitations and enhance the performance and lifetime of WSNs, we propose a new scheme based on the AODV (Ad hoc On-Demand Distance Vector) protocol within a mesh networking framework. Our approach achieves optimal results, including 100% throughput, minimal jitter, and significant improvements in network life-time.
Authors - Pranav Mittal, Prerna, Dhruv Bansal, Nikhil Panwar, Krish Tyagi Abstract - Stock market prediction is an intricate process in financial analysis, as its main aim is to predict the price trends and thus help to make trading strategies. But the stock market is so unpredictable with various factors affecting it, and thus predicting whether the value will rise or not becomes a difficult task. This study aims to predict the stock market prices with historical data that will be achieved via Deep Learning techniques in particular LSTM networks. Due to its strength in learning long-term dependencies and preserving the sequence information, LSTM (one of the variants of RNNs) is ideal for time-series data. Our approach leverages a dataset of daily stock prices from various financial indices over multiple years. The data is preprocessed using normalization techniques to improve model accuracy. The LSTM model is then compared to a traditional feed-forward neural network to demonstrate the superiority of LSTM in predicting shortterm stock trends. The models are optimized using the Adam optimizer, The results indicate that LSTM significantly outperforms the conventional models in forecasting accuracy. Moreover, this research introduced a hybrid LSTM-CNN method to extract features and prediction firmly. This study will contribute to financial forecasting by utilizing deep learning techniques and real trading scenarios. The research work was carried out using the programming language known as Python, deep learning tools such as TensorFlow and Keras, data management libraries such as Pandas and NumPy, and data representation software such as Matplotlib. It was found that the LSTM model has a much higher level of success when time series patterns are approximate than any other type of neural network, hence stock price predictions are more accurate. This study helps in how LSTM networks can be useful for forecasting in finance hence would be helpful to traders and market analysts.
Authors - Sheetal Phatangare, Komal Potdar, Yash Mahajan, Mandar Pandagale, Vivek Nikam Abstract - University students frequently encounter challenges in retrieving relevant academic information due to the limitations of traditional search engines. This research introduces KnowledgePilot, the first 1-bit Large Language Model (LLM) specifically designed to support university students. Leveraging the BitNet b1.58 architecture, which employs ternary parameterization (-1, 0, 1), KnowledgePilot achieves high performance with reduced computational costs, making it both resource-efficient and fast. The system integrates Retrieval Augmented Generation (RAG) pipelines, enabling it to access external academic data sources, thus minimizing hallucination issues common in LLMs and providing accurate, context-specific responses. The research also encompasses the development of tools for file conversion, dataset creation, model pretraining and fine tuning. Comprehensive evaluations will measure the system’s performance and user satisfaction, demonstrating its potential to significantly enhance student access to academic resources, while setting the stage for future advancements in low-bit AI technologies for education.
Authors - Yash Chavan, Arnav Sonawane, Arpit Pattiwar, Aditya Nagdive, Kaushalya Thopate Abstract - In today's fast-moving world, consumers rely on packaged foods. This is extremely important for easy access to detailed and personalized nutritional information. This project focuses on developing mobile applications for barcode scanning. This includes extensive food details, including ingredients, nutritional value, allergen warnings, and personalized consumption recommendations based on a person's health. Applications written with Python and Kivy provide a seamless user experience, allowing individuals to scan barcodes and upload images of ingredients to extract and analyze related information. Additionally, it includes optical character detection (OCR) using Tesseract, which extracts text from photos to allow users to analyze the ingredient list and nutritional name, even if barcode scans are not possible. By taking into account user nutritional limitations or illnesses such as diabetes, lactose intolerance, or gluten sensitivity, this application provides tailor-made health advice and helps individuals make found food decisions appropriately. A secure user authentication system improves the experience by storing your preferences and receiving recommendations created by tailors. The main goal of this project is to enable consumers to choose food in real time and promote healthier consumption habits. The combination of barcode scanning, OCR, and a structured database causes applications to close the gap between the complexity of food indicators and user understanding. Future improvements include mechanically learning-based ingredients, integration into real-time product databases, and expansion of several platforms beyond Android. This initiative represents an important step in using technology to improve consumer health awareness and ensure safer and sounder decisions for food consumption.
Thursday August 27, 2026 12:30pm - 2:30pm IST Virtual Room BGOA, India
Authors - Evangeline R C, Krupa Nirmal, Laasya P, Aishwarya K Abstract - Real-time pedestrian trajectory prediction is essential for enhancing safety and urban mobility, particularly in dense and dynamic environments. This paper introduces a video data processing system that accurately predicts pedestrian movement by analyzing sequences of video frames in real time. The system effectively handles challenges such as overlapping individuals, partial occlusions, and diverse walking behaviors, making it suitable for real-world deployment. The architecture is designed to be both modular and scalable, allowing for seamless integration into various applications such as traffic management, urban planning, and pedestrian safety enhancement. A user-friendly interface provides real-time visualization of the predicted trajectories, enabling accessibility for both technical and non-technical stakeholders, including urban planners and public safety officials. Extensive experiments conducted on multiple diverse datasets demonstrate the system’s reliability and accuracy across various conditions, including crowded scenes and irregular pedestrian movement. The system successfully captures complex behavior patterns and provides predictive insights that can help reduce pedestrian-related accidents. This research contributes significantly to the field of intelligent pedestrian monitoring systems. By combining real-time responsiveness with accurate trajectory prediction, the proposed system supports the development of smarter and safer urban infrastructure, fostering proactive decision-making and improved pedestrian safety in modern cities.
Authors - Nisha Dubey, Randeep Singh Abstract - Facial profile classification and acknowledgment have different applications in security, perception, and identity affirmation. This paper proposes a novel approach utilizing Convolutional Neural Frameworks (CNNs) to classify and recognize facial profiles. The proposed system utilizes a significant CNN designing to remove solid highlights from facial profiles, taken after by a classification layer to recognize profile classes (e.g., cleared out, right, frontal). The illustrate is ready on a tremendous dataset of facial profiles and finishes tall accuracy in classification (95.2%) and affirmation (92.5%) errands. Test comes around outline the system's quality to assortments in lighting, pose, and expression. Besides, the proposed system outflanks existing techniques in facial profile classification and affirmation. This work contributes to the movement of facial examination development, engaging its course of action in real-world applications such as identity affirmation, get to control, and perception. In particular, facial profiles provide an interesting challenge due to the distinctive variety of lighting, attitude, expression and disorders. Furthermore, large data records and accessibility of arithmetic violations have made it possible to prepare violent .CNN models for facial profile classification and detection
Authors - Meena Rani, Randeep Singh Abstract - The exponential growth of smart devices and advanced image editing tools has made detecting and localizing image forgeries critical for ensuring digital content integrity. This paper focuses on developing a robust and scalable model for passive image forgery detection using convolutional neural networks (CNNs). Leveraging datasets like CASIA1 and MICC-F220, the study aims to identify tampered regions in digital images by analysing noise patterns, pixel-level anomalies, and compression artefacts. The proposed methodology integrates preprocessing, model training, and validation using diverse datasets to enhance detection accuracy and scalability. Compared to traditional techniques, the deep learning-based approach shows significant improvements in detecting complex forgeries, including splicing and copy-move manipulations. Applications of this research extend to digital forensics, media authentication, and cybersecurity. The findings underscore Deep learning systems' show promise to tackle new issues in picture forgery detection and localization
Authors - Kritika Benjwal, Rishika Agrawal, Rashi Gupta, Dinesh Kumar Saini Abstract - The ublic distribution system (PDS) plays a vital role in eradicating hunger and ensuring food security across the world[1]. However, there are certain challenges like beneficiary identification, inconsistent transaction, diversion of grains during procurement at different stages to the open market, ration shop owners selling subsidized goods at higher prices and most importantly the paper focusses on supply chain leakages.[2] This paper proposes a conceptual model for implementing blockchain in PDS and eradicating all the supply chain leakages and making PDS fair[3]. It first focusses on operations of PDS, then assessing all the possible loopholes and implementing blockchain for removing these loopholes. It proposes the idea in which it leverages the benefits of using smart contract and consortium-based ecosystem that can bring efficiencies in PDS.
Authors - Sandhya Borkar, Shital Patil Abstract - The rapid development of the Internet of Things (IoT) has made the provision for innovative solutions in various sectors, including fuel management. The research presents the design and analysis of a smart IoT-based system for fuel dispensing, aimed at improving the efficiency, transparency and security of fuel distribution. The proposed system uses microcontrollers, sensors and real-time data communication technologies to automate fuel dispensing, monitor fuel levels and prevent theft or misuse. Key components include flow sensors to measure fuel output, RFID modules for secure user authentication and cloud-based platforms for remote monitoring and control. Additionally, mobile applications provide users with instant transaction records, fueling history and alerts. The system undergoes performance evaluation to ensure precise fuel measurement and seamless data synchronization. The results demonstrate significant improvements in operational efficiency and customer satisfaction, reducing manual errors, fuel theft and fraud, operational downtime, high maintenance and labor cost. This smart fuel dispensing system offers a scalable and cost-effective solution suitable for fuel stations, logistics companies and industrial applications, contributing towards smarter resource management and enhanced energy distribution practices.
Authors - Pravin Game, Shubham Bhingardive Abstract - Deepfake technology, which enables manipulation of images and videos, poses serious threat to media integrity and cyber security. Existing detection models often struggle with accuracy due to data complexity and variability. This study introduces a hybrid deepfake detection model that combines MobileNetV2, EfficientNetB7 and Vision Transformer (ViT) to enhance feature extraction and classification. ViT provides strong pattern recognition, EfficientNetB7 offers scalable accuracy and MobileNetV2 ensures lightweight processing. The model is trained on a publicly available datastet from Yonsei University, consisting of real and fake facial images. Techniques such as data augmentation and image resizing improve generalization. Experimental results show that the proposed model achieves 94.64% accuracy, 93.55% precision, 96.67% sensitivity, 92.31% specificity and 95.08% F1 score outperforming precious methods. These improvements ase statistically significant (p < 0.05). The results highlight the effectiveness of multi-model fusion for robust deepfake detection offering a scalable and reliable solution for applications in digital forensics, information verification and cybersecurity.
Authors - Pravin Game, Shubham Bhingardive Abstract - The correct identification of heart disease is essential for successful treatment and management. In this work, we assess different machine learning algorithms' predictive power for diagnosing heart disease. On the dataset, we employed the method of principal component analysis (PCA) to choose features, we got top 9 principal components out of 13 features. Then, applied the hippopotamus optimization algorithm on that 9 principal components then trained and tested the model on eight different algorithms: Bagging, Boosting, Naive Bayes, K - Nearest Neighbors (KNN), Random Forest, Decision Tree, Support Vector Machine (SVM), and Logistic Regression(LR). The algorithm’s accuracy ranged from 86.81% to 94.53%, The most accurate methods were SVM, KNN and random forest. These findings show that machine learning algorithms may be able to help with heart disease and focus on the need of choosing suitable algorithms for exact and trustworthy clinical decision-making. Future research will concentrate on using sophisticated on feature selection and ensemble learning strategies to further increase model accuracy.
Authors - Ganesh Shivaji Pise, A D Londhe, Hrushikesh Jaivant Joshi, Bhagwan Dinkar Thorat, Yashita Parikshit Mahalle, Pankaj Chandre Abstract - Data leakage poses a significant threat to modern security systems, leading to unauthorized access and privacy breaches. Deep learning models have shown promise in detecting such anomalies; however, their black-box nature raises concerns regarding trust and interpretability. This paper explores the role of Explainable AI (XAI) in enhancing transparency and trust in deep learning-based data leakage detection. Various explainability techniques, including SHAP, LIME, and Grad-CAM, are integrated into a security framework to provide interpretability while maintaining detection accuracy. The proposed architecture bridges the gap between AI-driven security solutions and human decision-making, enabling security analysts and compliance officers to make informed assessments. Additionally, the study evaluates different XAI approaches based on accuracy, interpretability, and scalability to identify optimal techniques for real-world security applications. The findings highlight the importance of balancing explainability with performance to ensure robust and trustworthy cybersecurity solutions.
Authors - Priya Surana, Soham Jadhav, Janhvi Jathot, Arnav Joshi, Roshani Kadam Abstract - Landslides are natural disasters posing great risks to life, infrastructure, and the environment. Timely and accurate predictions are highly beneficial to reduce such impact. The advent of machine learning (ML) and deep learning (DL) has significantly improved the state of landslide prediction models. The review outlines the various ML and DL techniques adopted for landslide prediction and gives a brief account of methodologies, applications, benefits, and limitations. This is mainly the melding of ML and DL techniques, such as Random Forest (RF), Support Vector Machines (SVM), Convolutional Neural Networks (CNN), and Long Short-Term Memory (LSTM) networks, for the enhancement of the predictive ability of such models. Key challenges in landslide prediction, such as data availability, model interpretability, and computational complexity, alongside future directions, will be discussed to contribute to the robustness of landslide prediction models. Finally, inferences will be drawn as to the significance of hybrid ML-DL approaches in pushing forth landslide prediction models into better accuracies and reliability (Khuc, T.D., et al, 2023)(Wu, X., et al, 2023).
Thursday August 27, 2026 3:30pm - 5:30pm IST Virtual Room BGOA, India
Authors - A.Punidha, E.Arul, E.Yuvarani, S.Rajasakaran Abstract - With the rapid expansion of Internet of Things (IoT) and smart device ecosystems, security threats such as malware attacks have become a critical concern. Traditional signature-based malware detection methods struggle to detect evolving and polymorphic threats, necessitating the development of intelligent, data-driven cybersecurity mechanisms. This study proposes a novel malware detection framework that integrates optimized feature engineering and deep neural networks (DNNs) to classify malware in smart devices with high precision. The approach focuses on behavioral feature extraction, including API call sequences, network activity logs, and application permissions, followed by feature selection techniques to reduce dimensionality while retaining key discriminative attributes. A comparative analysis of various machine learning (ML) models, including Random Forest, Support Vector Machine (SVM), and Deep Learning models, demonstrates that the proposed feature engineering-enhanced DNN model achieves 96.1% accuracy, outperforming conventional methods. Extensive experimentation on a real-world dataset of 10,000 smart device applications showcases the robustness and scalability of our framework. This research contributes to enhancing security in smart environments by providing an adaptive and computationally efficient malware detection system.
Thursday August 27, 2026 3:30pm - 5:30pm IST Virtual Room BGOA, India
Authors - Jolou Vincent M. Jala, Everly A. Nacalaban, Nenon Roy A. Sandinao, Ryan Boyd D. Origines, Randy Joy M. Ventayen, Neilson D. Bation Abstract - Artificial intelligence (AI) has completely transformed enterprises and organizations all over the world due to its propensity to spur innovation and mimic operational efficiency (Jala, J.V.M. et al., 2024). Through its capability to boost learning outcomes, promote inclusion, and streamline operations, artificial intelligence (AI) is revolutionizing higher education. This study investigates the influence of artificial intelligence in higher education in the Philippines. The study specifically seeks to understand how artificial intelligence (AI) can be used in higher education institution in terms of personalized learning amidst large class sizes, access to education in rural areas, solving job skills mismatch, modernizing administrative procedures in inadequate resources institutions, artificial intelligence powered innovation and research, challenges of embracing artificial intelligence such as digital literacy and infrastructure, social and ethical implications and resistance to change and faculty development Moreover, this work also examines the ethical considerations in employing arti-ficial intelligence in higher education in the Philippines, precisely in terms of security and data privacy, fairness and algorithmic bias, digital divide, human supervision and accountability, autonomy and consent, influence on staff and faculty roles and intellectual property and academic integrity. To realize this objective, the proponents essentially examined 170 publications in the literature that were indexed by Scopus to look at artificial intelligence in the context of higher education. This finding highlights artificial intelligence’s essential role in embracing challenges and improving higher education in the Philippines while emphasizing ethical considerations such as fairness and data privacy. (Jala, D.J.V., 2025).
Authors - Kishan Raj, Samyuktha Vimal, V Shini Abstract - In recent years, the Indian online fashion e-commerce industry has experienced significant transformations due to fast-paced technology, consumer behavior changes, and a growing e-commerce landscape. Competition is fierce, especially in the e-lifestyle market as its payoff will exceed $30 billion by 2025, developing brand equity understandably becomes a key factor in maintaining long-term success. It is understood that brand equity is a significant factor as it directly affects consumers' trust, buying decisions, and consumer retention, which are also vital elements in preserving brand equity in an industry where distinctions matter. This study's goal is to examine the impact of marketing mix variables 7Ps including Personalization, which has just recently emerged as a particular critical factor in the e-commerce business, on brand equity in the Indian online fashion e-commerce industry. Quantitative research methods were adopted to analyze data provided by consumers regarding the influence of such factors. The analyses indicate that Personalization and Place affect brand equity substantially, whereas traditional elements such as Price and Promotion do not exert such influence in online fashion retailing. Indian consumers seem increasingly to make a shift toward a digital-first shopping experience, the marketing strategies of brands must follow suit by creating engaging, personalized, and innovative interactions. These research findings will provide some strategic recommendations to online fashion retailers, marketers, and industry gurus seeking to consolidate their brand positioning in this ever-evolving and highly competitive marketplace.
