Daily Archives: August 6, 2026

Authors: Arhaan Shaikh

Abstract: Parking congestion has become one of the major challenges in modern urban areas because of rapid population growth, expansion of cities, and the continuously increasing number of private vehicles. In many commercial zones, res-idential complexes, shopping malls, railway stations, airports, and educational campuses, drivers often face difficulty in finding available parking spaces. This leads to unnecessary delays, traffic buildup, fuel wastage, driver frustration, and increased air pollution. Traditional parking management systems generally depend on fixed thresholds or simple binary decision-making methods, where congestion is classified only as full or empty. Such systems are not flexible enough to handle real-time changes in parking demand and uncertain traffic situations. This paper presents a Mamdani fuzzy logic-based smart parking congestion detection system that can intelligently es-timate parking congestion levels using two important input parameters: vehicle count and free slot availability. Instead of using rigid boundaries, fuzzy logic uses linguistic terms such as Low, Medium, and High to represent real-world conditions more naturally. The proposed model uses triangular membership functions for fuzzification, a nine-rule inference engine for decision-making, and centroid defuzzification to generate a final congestion output. The system provides smoother transitions between congestion states, better handling of boundary values, and more realistic results compared to conventional methods. Due to its low computational complexity, the proposed system is highly suitable for real-time embedded devices, IoT-based smart city applications, and automated parking guidance systems.

Published by: vikaspatanker

Authors: Moaiz Kazi, Sufiyan Ansari, Arhaan Shaikh, Madhvi Saxena, Usaid Khairdi

Abstract: The digital world we live in today is creating an amount of unorganized data. This has led to something called hoarding. People who use cloud storage often feel overwhelmed by the number of files they have. They have a time searching for things and their organized systems start to fall apart. To solve this problem we are introducing XAI-CloudAssist. XAICloudAssist is a tool that is designed to work in the cloud. It automatically sorts files into categories using a group of Random Forest models. Unlike automated systems that are hard to understand our approach is transparent. We use something called Explainable AI to make sure people can see how it works. We use SHAP values to show why each file is sorted into a category. We give a score to each piece of information about the file to show how important it is. When we tested XAI-CloudAssist it was able to sort files 98 percent of the time. This shows that things like how space a file takes up and how often it is used are very good, at predicting what category it belongs in. XAICloudAssist also explains why each file is sorted into a category. This helps people understand the decisions it makes and trust that it is working correctly. XAI-CloudAssist is related to Cloud Computing and File Organization and Machine Learning and Explainable AI and SHAP and Random Forest and Automation and Data Governance.

Published by: vikaspatanker

Authors: Moaiz Kazi, Sufiyan Ansari, Arhaan Shaikh, Usaid Khairdi

Abstract: The modern digital world is generating unprece-dented amounts of unstructured data, which has given rise to “digital hoarding.” Cloud users often feel overwhelmed by massive file repositories, experience slow searches, and see their organized systems break down. To tackle this, we introduce XAI-CloudAssist, a robust, cloud-native assistant designed to automatically classify files using an ensemble of Random Forest models. Unlike traditional automated systems that operate as black boxes, our approach puts transparency first by integrating Explainable AI. We use SHAP values to reveal the reasoning behind each classification, assigning quantitative importance scores to metadata features. In experiments, the system achieved 98 Percentage classification accuracy, showing that metadata-driven cues like storage footprint and access frequency are highly predictive. Moreover, XAI-CloudAssist provides localized expla-nations for every categorization, helping users understand the de-cisions and fostering trust in cloud automation. Keywords: Cloud Computing, File Organization, Machine Learning, Explainable AI, SHAP, Random Forest, Automation, Data Governance.

Published by: vikaspatanker

Authors: Balvir Singh Thakur, Chakshu Bhardwaj

Abstract: Artificial intelligence (AI) is frequently suggested as a means to reduce the participation thresholds that presently inhibit gamers with disabilities from playing digital games. However, evidence to support this assertion has been scattered across academic literature, open datasets, industry standards, and practitioner discourse and seldom consolidated. This review brings together and compares the four kinds of evidence, not to develop new theoretical concepts or to present new experimental data, but to emphasize the areas of agreement and disagreement in the literature. Following the PRISMA 2020 methodology, a repeatable search and filtering pipeline was developed, covering IEEE Xplore, the ACM Digital Library, Scopus, Web of Science, and Google Scholar, for the period 2018-February 2026, further enriched by a qualitative evidence synthesis of open accessibility resources: large-scale Steam review datasets, the Game Accessibility Guidelines (GaG), the IncluSet repository, and AbleGamers Accessible Player Experiences (APX), and a comparative analysis of academic research results against player signals and practitioner recommendations. Analysis of this corpus reveals recurrent themes, with advanced AI applications being largely limited to speech-to-text captioning, text-to-speech audio, computer vision for navigation and object recognition, and reinforcement learning for adaptive difficulty. Cutting-edge yet less-explored solutions involve large language models and other forms of generative AI. There is considerable agreement between literature, expressed user interest, and practitioner recommendations regarding captioning, data privacy in adaptive accessibility systems, and multiplayer game balance. Still more research on making games accessible to players with disabilities is focused on visual and hearing impairments than on motor or cognitive disabilities. Together, these point to opportunities for an integrated agenda that prioritizes disability-balanced training data, ecologically valid evaluation of gaming accessibility, and participatory AI design approaches centered on user privacy.

DOI: https://doi.org/10.5281/zenodo.21822669

Published by: vikaspatanker

Authors: Thogaru Aahalya, P.S.V Krishna

Abstract: The widespread adoption of social media platforms has transformed the way disaster-related information is generated and disseminated, providing valuable real-time insights during emergency situations. However, extracting reliable and actionable information from the enormous volume of unstructured social media content remains a significant challenge due to data noise, misinformation, and incomplete contextual information. This paper presents an intelligent disaster monitoring framework that integrates artificial intelligence, natural language processing, sentiment analysis, and location intelligence to enable real-time detection of disaster events from social media streams. The proposed framework systematically collects and preprocesses user-generated content, identifies disaster-related posts through machine learning techniques, extracts geographic information to determine affected regions, and evaluates public sentiment to assess the severity and urgency of ongoing incidents. By combining textual analysis with geospatial intelligence, the system enhances situational awareness and provides timely decision support for emergency response agencies. Experimental evaluation demonstrates that the proposed framework achieves reliable disaster detection performance while improving the accuracy of event localization and public sentiment assessment. The proposed approach offers a scalable and intelligent solution for disaster management, facilitating faster emergency response, effective resource allocation, and improved public safety during crisis situations.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue4.152

Published by: vikaspatanker

Authors: Reddi Sri Venkata Lakshmi Lahari, K Ravi Kumar

Abstract: The increasing volatility of airline ticket pricing, driven by dynamic market demand, seasonal variations, operational costs, and competitive pricing strategies, has made accurate flight fare prediction a challenging task. Conventional forecasting approaches often struggle to model the nonlinear and time-dependent relationships that influence airfare fluctuations. This paper presents an intelligent web-based flight price prediction framework that integrates advanced machine learning and deep learning techniques to deliver accurate and real-time airfare forecasts. The proposed system employs comprehensive data preprocessing, feature engineering, and sequential learning to capture complex pricing patterns from historical flight data. A Long Short-Term Memory (LSTM) network is utilized to model temporal dependencies, while adaptive learning mechanisms enable the framework to remain responsive to continuously changing market conditions. The developed web application provides users with an interactive platform for obtaining instant fare predictions, thereby supporting informed travel planning and strategic pricing decisions. Experimental evaluation demonstrates that the proposed framework achieves reliable predictive performance and effectively models dynamic airfare trends. The proposed solution offers a scalable, data-driven decision support system for airlines, travel agencies, and passengers, contributing to the advancement of intelligent transportation analytics and real-time predictive systems.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue4.151

Published by: vikaspatanker

Dr. K. Murali Mohan Achari
Affiliation	Assistant Professor,Department of Chemistry, Sri Venkateswara College, University of Delhi, New Delhi.
Email-Id:	mmakamsali@svc.ac.in
Publication: . Books: Kamsali, M. M. A., & Thoti, V. (2025). Biomedical applications of functionalized magnetic nanomaterials (FMNs). In Nanostructure Science and Technology (pp. 39–58). doi:10.1007/978-3-031-97199-0_2. Publications: Seema, V., Kamsali, M. M. A., Alakonda, L. M., Varala, R., Hussein, M., & Alam, M. M. (2026). Mini-update on the applications of Hypophosphites in organic synthesis with a special focus on sodium hypophosphite (NaH2PO2). Tetrahedron Letters, 174(155875), 155875. doi:10.1016/j.tetlet.2025.155875. Majumder, A. B., Kamsali, M. M. A., Varala, R., & Ansari, S. A. (2026). Biocatalysis in bioorthogonal reactions: Use of hydrolases and transferases for selective modifications. Mini-Reviews in Organic Chemistry, 23(1), 10–19. doi:10.2174/0118756298358967250117114300. Alakonda, L. M., Kamsali, M. M. A., Alam, M. M., & Varala, R. (2026). Recent advances in ammonium iodide (NH 4 I)‐mediated transformations: Toward greener and metal‐free organic synthesis. European Journal of Organic Chemistry, (e202501115). doi:10.1002/ejoc.202501115. Varala, R., Seema, V., Kamsali, M. M. A., Kousar, N., Hussein, M., & Alam, M. M. (2026). Visible light-driven innovative approaches for the synthesis of quinoline and isoquinoline based heterocycles. Tetrahedron Letters, 175(155922), 155922. doi:10.1016/j.tetlet.2025.155922. Varala, R., Kamsali, M. M. A., Hussein, M., & Alam, M. M. (2025). Applications of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) in heterocyclic ring formations. Organic & Biomolecular Chemistry, 23(41), 9285–9329. doi:10.1039/d5ob01209k.

Published by: vikaspatanker

Authors: Rabins Porwal

Abstract: — Deep convolutional networks now match or exceed specialist performance on several image-based diagnostic tasks, yet their adoption in clinical practice remains limited by a problem that accuracy alone cannot solve: a clinician asked to act on a prediction cannot see why it was made. Post-hoc saliency methods offer a partial answer, but different methods applied to the same network routinely disagree, and a map that looks convincing is not necessarily one that reflects the computation the model actually performed. This paper proposes an Explainable Deep Learning (XDL) framework in which interpretability is a training objective rather than an afterthought. An attention-guided refinement module reweights backbone feature maps so that spatial evidence is concentrated before classification. Three complementary attribution methods – Grad-CAM++, integrated gradients and GradientSHAP – are then fused into a single saliency map, and the disagreement among them is quantified as an explanation-consistency score. Finally, a deletion-based faithfulness penalty is added to the loss, so that the network is optimised not only to classify correctly but to concentrate its evidence on regions whose removal genuinely changes the prediction. Evaluation across five public datasets spanning radiography, dermoscopy, fundus photography, magnetic resonance imaging and histopathology gave a mean accuracy of 93.7 per cent, 2.1 percentage points above the strongest baseline. More importantly, faithfulness improved substantially: deletion AUC fell from 0.157 to 0.128 and agreement with expert-annotated lesion masks rose from 0.452 to 0.518 in intersection over union. A blinded review by three clinicians rated the fused maps 4.2 out of 5 for plausibility against 3.6 for the best competing method.

DOI: https://doi.org/10.5281/zenodo.21819333

Published by: vikaspatanker

Authors: Sakthi Sankara Balaji Sathyamurthy

Abstract: Wholesale banking credit decisions involve large and interconnected exposures, evolving borrower conditions, collateral movements, covenant compliance, and changing market environments. Conventional credit-risk models are commonly designed for one-time prediction tasks and may not adequately support sequential decisions such as adjusting exposure limits, revising pricing, requesting additional collateral, or escalating cases for review. This study proposes an explainable reinforcement learning framework for real-time credit-risk decisioning in wholesale banking ecosystems. The framework represents credit management as a sequential decision problem in which an agent observes dynamic borrower, facility, portfolio, and market conditions before recommending appropriate risk actions. It integrates offline reinforcement learning with policy constraints to reduce the risk of unsafe or non-compliant recommendations in regulated banking environments. The proposed model incorporates an explainability layer that provides feature-based explanations, reward decomposition, policy rationale, and counterfactual action analysis. These outputs are intended to clarify why a specific credit action is recommended, which risk factors influenced the decision, how the recommendation aligns with risk appetite, and what changes in borrower conditions could lead to an alternative outcome. The study adopts a design science and empirical evaluation approach using historical wholesale credit data, with performance assessed against conventional statistical models, machine- learning models, rule-based decision engines, and non-explainable reinforcement-learning approaches. Evaluation criteria include credit-risk performance, expected-loss control, risk-adjusted return, decision latency, policy compliance, robustness, and explanation quality. The study contributes a structured framework for applying explainable reinforcement learning to high-impact credit decisions. It also provides guidance for controlled deployment through offline training, human oversight, model validation, audit logging, and continuous monitoring. The proposed approach may strengthen the timeliness, consistency, transparency, and governance of credit-risk decisioning across wholesale banking operations.

DOI: https://doi.org/10.5281/zenodo.21819118

Published by: vikaspatanker