Manuscripts

Recent Papers

Research Paper

Cab Fare Prediction Machine Learning Model

Predicting cab fares accurately is crucial for urban transportation, benefiting both passengers and service providers. This research explores machine learning techniques to enhance fare prediction using real-world trip data. Various models, including Linear Regression, Decision Trees, Random Forest, Gradient Boosting, and XGBoost, were evaluated. The Gradient Boosting Regressor emerged as the best-performing model after hyperparameter tuning, achieving high prediction accuracy. The study highlights the significance of trip distance and pickup time in fare estimation. Future enhancements include integrating weather data and deploying the model as a real-time API service to improve usability and precision.

Published by: Purvank Chauhan, Shubham Upadhyay

Author: Purvank Chauhan

Paper ID: V11I1-1480

Paper Status: published

Published: March 27, 2025

Full Details
Research Paper

Reinforced War Bunker Construction

Propagation of shock waves in partially- or fully-confined environments is a complex phenomenon due to the possibility of multiple reflections, diffraction and superposition of waves. In a military context, the study of such phenomena is of extreme relevance to the evaluation of protection systems, such as survival containers, for personnel and equipment. True scale testing of such structures is costly and time consuming but small-scale models in combination with the Hopkinson- Cran scaling laws are a viable alternative. This paper combines the use of a small-scale model of a compound survival container with finite element analysis (with LS- DYNA) to develop and validate a numerical model of the blast wave propagation. The first part of the study details the experimental set-up, consisting of a small-scale model of a survival container, which is loaded by the detonation of a scaled explosive charge. The pressure-time histories are recorded in several locations of the model. The second part of the study presents the numerical results and a comparison with the experimental data.

Published by: Aryan Sable, Priyanshu Arde, Siddharth Patil, Lakshmi Hanchate, Sagar Mungase

Author: Aryan Sable

Paper ID: V11I1-1472

Paper Status: published

Published: March 27, 2025

Full Details
Survey Report

Need for Privacy-Preserving AI for Secure Data Sharing in Cybersecurity

The purpose of this exploratory study is to look into the necessity for Privacy-Preserving Artificial Intelligence (AI) in secure data sharing in the context of cybersecurity. The research design includes a comprehensive examination of the current literature and a survey questionnaire with industry professionals. The findings show a growing demand for privacy-preserving AI solutions in cybersecurity, driven by increased data privacy rules and the escalation of data breaches. The study found that typical data-sharing mechanisms frequently reveal sensitive information, rendering them inappropriate for handling secret data. The practical ramifications of these findings are substantial. They highlight the importance of enterprises implementing privacy-preserving AI solutions to improve data security while adhering to privacy standards. Such solutions can assist firms in leveraging their data for insights while maintaining the privacy of individuals' information. However, the study does identify shortcomings. The adoption of privacy-preserving AI systems can be difficult due to their computational cost and the potential decrease in data value caused by extra noise for privacy preservation. Furthermore, a lack of awareness and comprehension of these solutions among businesses creates additional hurdles to their implementation. The study underlines the critical need for Privacy-Preserving AI for secure data exchange in cybersecurity and advocates for increased awareness and research in this area to address the stated problems.

Published by: Tejas Yeole, Abhinita Daiya

Author: Tejas Yeole

Paper ID: V11I1-1460

Paper Status: published

Published: March 27, 2025

Full Details
Research Paper

Analysis of CNN Models for Melanoma Detection

Melanoma is the deadliest type of skin cancer that needs to be detected at its early stages to prevent fatality. Using dermoscopy images of the lesion a computer-based system trained with deep learning will be developed to detect melanoma. The model will identify and categorize melanoma with intricate image processing and classification algorithms, which will be trained on a labeled dataset. Some of the goals of this project are to compile and preprocess a dataset of dermoscopy images labeled with benign lesions and melanoma, evaluate using metrics such as AUC-ROC, accuracy and validation with external datasets, addressing bias while following clinical guidelines. At the end of this research, we hope to improve patient outcomes and lessen the cost of healthcare, making it affordable as well as increasing diagnostic accuracy, decreasing false positives, and assisting dermatologists in the early detection of the disease.

Published by: Adithya.R, Mohammed Yassin A, Dr Sonia Jenifer Rayen

Author: Adithya.R

Paper ID: V11I1-1444

Paper Status: published

Published: March 27, 2025

Full Details
Research Paper

Differentiating Fault Current from Leakage Current during IC Testing

Differentiating Fault Current from Leakage Current During IC Testing As integrated circuit technology advances, the intricacies related to fault identification and leakage current evaluation have increased markedly. Although conventional I_DDQ (quiescent supply current) testing protocols exhibit effectiveness in detecting major defects, they often struggle to distinguish between typical leakage currents and those reflective of genuine faults, thereby prolonging testing durations. Consequently, current sensors typically initiate measurements once transitions are finalized. In this investigation, we utilize a simulation technique to corroborate the effectiveness of an innovative methodology articulated in [1], which can be used to improve the throughput of the current testing process by detecting the faults using AC components of the current, thereby overcoming a constraint of traditional methods.

Published by: Yasser A. Ahmed

Author: Yasser A. Ahmed

Paper ID: V11I1-1488

Paper Status: published

Published: March 27, 2025

Full Details
Research Paper

2D Platformer Game Development with Godot Engine

This paper explores the development of a 2D platformer using the Godot Engine, emphasizing its node-based design and procedural generation. By adopting an agile workflow, the project achieved stable 60 FPS on mid-tier hardware, with user tests (n=50) highlighting responsive controls (93% approval) and dynamic levels. Godot’s efficiency and open-source flexibility proved ideal for indie teams, though advanced debugging tools were limited. Findings affirm its viability for cost-effective, engaging 2D game development.

Published by: Neeraj Pradeep Bharambe, Yash Prasad Mhaddalkar, Harsh Manohar Yeram, Vidhi Santosh Jadhav

Author: Neeraj Pradeep Bharambe

Paper ID: V11I1-1501

Paper Status: published

Published: March 27, 2025

Full Details