Daily Archives: June 25, 2026

Authors: Gitesh Patil, Sakshi Mahajan, Shloka Shetty, Samruddhi Nevse, Dr. Arati R. Deshpande

Abstract: Conversational AI companions operating in emo-tionally sensitive and therapeutic contexts require the joint integration of speech understanding, affective reasoning, and photorealistic visual feedback — capabilities that existing systems address only in isolation, and largely through cloud-dependent infrastructure that introduces recurring costs and privacy con-cerns. Although recent advances in large language models and neural speech synthesis have improved the quality of automated dialogue, current systems lack the structural coupling between emotion recognition, semantic memory, and avatar-driven fa-cial expressiveness necessary for naturalistic human-computer interaction. This paper presents EmotionSync, a locally-hosted conversational AI companion capable of performing end-to-end affective interaction while maintaining real-time responsiveness. The proposed system integrates faster-Whisper-based speech-to-text transcription, Wav2Vec2 speech emotion recognition, retrieval-augmented generation over a ChromaDB vector store, locally-served LLaMA 3.1 language model inference, Microsoft Edge neural text-to-speech synthesis, and NVIDIA Audio2Face 3D blendshape-driven avatar animation within a unified Web-Socket streaming pipeline. By enforcing phrase-boundary audio chunking and performance.now()-anchored blendshape dispatch, the framework ensures frame-accurate lip synchronization and emotionally coherent response generation. The proposed frame-work contributes toward practical, privacy-preserving affective AI companions suitable for therapeutic, educational, and social interaction applications.

Published by: vikaspatanker

Authors: Vishal Paithankar, Shekhar Kale

Abstract: Ultra High Performance Concrete (UHPC) is known for its exceptional mechanical strength and durability, yet its effectiveness is notably compromised when subjected to high temperatures. This study investigates the temperature- dependent mechanical properties of UHPC, with emphasis on residual compressive and tensile strengths. The influence of critical parameters such as steel fiber content, polypropylene fiber dosage, water-to-binder ratio, and supplementary cementitious materials is systematically analyzed. Experimental findings from existing literature indicate that increasing temperature leads to strength degradation due to microcracking, matrix densification loss, and fiber–matrix debonding. The inclusion of polypropylene fibers is found to mitigate explosive spalling by enhancing vapor pressure release. Furthermore, artificial neural network models are explored to predict residual mechanical properties of UHPC under thermal exposure. The outcomes contribute to a better understanding of UHPC behavior in fire-prone structural applications.

DOI: http://doi.org/10.5281/zenodo.20855989

Published by: vikaspatanker

Authors: Abhijit Madhavrao Thakare, Chaitanya Mishra, Abhijit M. Thakare

Abstract: Bridges are essential components of transportation infrastructure, requiring effective inspection and management to ensure safety and durability. This study presents a comparative analysis of bridge inspection practices and Bridge Management Systems (BMS) in India, the United States, and China. It evaluates key aspects such as inspection methodologies, rating systems, inspector qualifications, and the use of advanced technologies. The results indicate that the United States follows a standardized and data-driven approach, India adopts a centralized and evolving system, and China demonstrates a technology-driven framework with real-time monitoring and predictive maintenance. The study highlights the growing role of advanced tools such as Structural Health Monitoring and non-destructive testing in improving inspection efficiency. It concludes that adopting modern technologies and standardized practices can significantly enhance bridge management systems.

DOI: http://doi.org/10.5281/zenodo.20855894

Published by: vikaspatanker

Authors: Sanjay B. Amrutkar, Dr. Dolly Thankanchan

Abstract: Transformer energization is commonly accompanied by severe inrush currents that may lead to protection maloperation, thermal stress, and power quality degradation. This paper presents a comprehensive comparative investigation of transformer inrush current mitigation using voltage ramping and closed-loop flux linkage control strategies. A nonlinear transformer model incorporating magnetic saturation and core losses is developed to evaluate peak inrush current, inrush ratio, thermal stress expressed through the i^2 t index, and control effort under multiple energization conditions. Simulation results demonstrate that the uncompensated case exhibits a peak inrush current of 83.37 A, corresponding to an inrush ratio of 16.47 and significant thermal stress. Voltage ramping effectively limits the peak inrush current to 7.15 A, achieving an inrush ratio of 1.41 and reducing the i^2 t energy by approximately 97%. The flux control strategy, while requiring higher injected voltage and control energy, maintains inrush currents below 16.3 A under ideal conditions and demonstrates strong robustness against residual flux and unfavorable switching angles, with peak inrush currents of 14.69 A and 7.94 A, respectively. Total harmonic distortion values approach 100% for all cases due to the non-periodic and transient nature of inrush current, indicating that THD is not a reliable metric during transformer energization. The results highlight the trade-off between mitigation effectiveness and control effort, and confirm the superior robustness of flux-based control under practical energization uncertainties.

DOI: http://doi.org/10.5281/zenodo.20855803

Published by: vikaspatanker

Authors: Mr.Sachin .S.Bhosale, Dr. Anand Singh Rajawat, Dr.P.R.Bhaldare

Abstract: This study demonstrate the hybrid framework model that combine Machine Learning (ML) and Deep Learning (DL) techniques for the detection of brain tumor on MRI scan dataset. We employ MobileNetV3 for deep feature extraction via transfer learning, followed by classification using Logistic Regression (LR), SVM, Random Forest, KNN, and XGBoost. Experimental results demonstrate that Logistic Regression paired with MobileNet features achieved superior performance (Accuracy: 95.02%, Precision: 94.78%, Recall: 94.53%, F1-score: 94.58%), outperforming more complex classifiers. This indicates that MobileNet-derived features create a nearly linearly separable representation, positioning LR as an efficient and effective tool for automated, accurate brain tumor diagnosis, thereby augmenting clinical decision-making.

DOI: http://doi.org/10.5281/zenodo.20855379

Published by: vikaspatanker

Authors: Ms. Sneha Dattatray Pekhale, Dr. Jatin Majithia

Abstract: Traditional integer-order differential equations are increasingly recognized for their limitations in capturing the non-local, memory-dependent and hereditary properties inherent in complex biological and physical systems. This review provides a comprehensive synthesis of recent research into the application of fractional-order derivatives—including the standard Caputo, Caputo-Fabrizio, Atangana-Baleanu and generalized ψ-Caputo operators—across diverse scientific domains. In epidemiology, these models have proven superior to classical approaches for analyzing the transmission dynamics of diseases such as Tuberculosis, COVID-19 and Dengue fever with some models achieving a 28.6% reduction in predictive error by accounting for specific population behaviors and environmental factors. In oncology, fractional modeling has refined the simulation of radiotherapy and chemotherapy by integrating vital radiobiological factors like cell repair and repopulation, leading to more precise treatment protocols. Beyond medicine, the sources demonstrate the utility of fractional calculus in modeling ecological food chain interactions, world population growth and USA GDP rates as well as optimizing multi-agent systems and gradient descent algorithms. By employing rigorous qualitative analyses (e.g.fixed-point theory) and advanced numerical schemes (e.g.Adams-Bash forth-Moulton method), these studies establish that fractional-order derivatives provide a more flexible and realistic framework for capturing the complexities of real-world phenomena. This review underscores the transformative potential of fractional calculus in enhancing predictive accuracy for public health management and socio-economic forecasting.

DOI: http://doi.org/10.5281/zenodo.20854700

Published by: vikaspatanker

Authors: Vishal Rathod, B. Rohith Patel, Deepika Borgoankar

Abstract: The increasing strain on healthcare systems across the globe has made patient triage an essential procedure to guarantee prompt and efficient medical care.Conventional man- ual triage techniques are constrained by concerns with scale, subjectivity, and human availability.Intelligent, automated sys- tems that can help with early patient triage have been made possible by recent developments in conversational AI and Natu- ral Language Processing (NLP).The development of NLP-based healthcare chatbots for patient triage is covered in this review paper, with a focus on ethical design, safety, and technological robustness.These systems may be able to comprehend natural symptom descriptions, offer non-diagnostic therapy recommen- dations, and improve access to healthcare by combining frame- works like Rasa with transformer-based language models (BERT, DistilBERT).The article analyzes previous research, discusses current research trends and limits, and investigates future options for implementing conversational triage systems that are safe, intelligible, and context-aware.

DOI: http://doi.org/10.5281/zenodo.20849583

Published by: vikaspatanker

Authors: Sonali Vidhate, Fuldeore Pritee, Jagtap Vaishnavi, Khairnar Vishakha, Aruba Kudai, Safa Madoo

Abstract: The rapid expansion of cloud computing, artificial intelligence (AI), and data-intensive applications has significantly increased the energy consumption of data centers, making sustainability a critical concern. Conventional energy optimization techniques such as virtualization, Dynamic Voltage and Frequency Scaling (DVFS), and static cooling mechanisms provide limited adaptability to modern, dynamic workloads. This research paper presents a comprehensive analysis of AI-driven green computing approaches for improving energy efficiency in data centers. Using insights from existing literature, this work proposes an intelligent framework that integrates machine learning, reinforcement learning, and predictive analytics to optimize workload distribution, cooling systems, and energy demand forecasting in real time. The proposed approach aims to reduce energy consumption, minimize carbon emissions, and improve Power Usage Effectiveness (PUE) while maintaining system performance. Additionally, novel innovations such as carbon-aware scheduling and renewable-energy-aware AI optimization are discussed to enhance sustainability. The findings indicate that AI-based energy management can achieve significant energy savings and support the development of future-ready green data centers.

DOI: http://doi.org/10.5281/zenodo.20849513

Published by: vikaspatanker

Authors: Dr. Harsha Sammangi, Poloju Pravalika

Abstract: Explainable artificial intelligence (XAI) is widely promoted as a remedy for algorithmic opacity, premised on the assumption that revealing a model's reasoning improves human oversight and decision quality. This study investigates the Explainable AI Paradox: the possibility that explanations simultaneously increase trust, reliance, and adoption while also producing overconfidence in flawed models — improving decision quality when models are valid but amplifying errors when models are not. Using a controlled experiment with 921 managers making pricing, lending, or inventory decisions assisted by AI systems of deliberately varied validity, participants were randomly assigned to one of five explanation conditions: no explanation, feature-importance, counterfactual, uncertainty-aware, or a combined feature-importance-plus-uncertainty condition. The study measured decision accuracy, confidence calibration, reliance behavior, override justification quality, and simulated financial outcomes. Results show that feature-importance explanations increased reliance (+9.8 percentage points, p < .001) and confidence (+0.61 scale points, p < .001) relative to no explanation, but produced the largest overconfidence increase (+0.084, p < .001) and the only negative financial outcome effect (–0.14 SD, p < .01) among all explanation types — concentrated specifically in the flawed-model conditions, where a significant Explanation × Model-Quality interaction (β = 0.047 to 0.089 across outcomes, all p < .001) confirms that feature-importance explanations' benefits accrue under valid models while their overconfidence costs accrue under flawed models. Uncertainty-aware explanations, by contrast, improved calibration (–0.058, p < .001), reduced overconfidence (–0.047, p < .001), and produced the only significant positive financial outcome (+0.29 SD, p < .001) relative to no explanation. A twelve-stage design intervention pilot demonstrates that combining five calibration-oriented design principles — feature reliability tagging, confidence-first ordering, disagreement prompts, active verification nudges, and explanation-accuracy feedback — reduces the Overconfidence Index by 83% (from 0.084 to 0.014, p < .001) relative to unmanaged feature-importance explanations. Thematic analysis of 40 participant interviews identifies six mechanisms underlying these patterns, including a 'plausibility heuristic substitution' through which surface-level explanation coherence substitutes for independent verification. The paper contributes a theory of the Explainable AI Paradox to behavioral information systems research, identifies model-quality and explanation-type interactions as the central moderating mechanism, and provides a five-level maturity roadmap and design decision framework for deploying explainable, uncertainty-aware managerial AI systems.

DOI: http://doi.org/10.5281/zenodo.20843984

Published by: vikaspatanker

Authors: Atul Dubey, Vikash, Ashish Juneja

Abstract: The paper is a study on the use of ISO shipping containers to build sustainable Urban Economical weaker section housing society for Slum living in Metropolitan cities in India. As per world bank report nearly 35.2 % of urban population of India living in Slums (2018). In Japan, China and Vietnam the use of container housing solved their housing problems. Shigeru Ban , a house hold name in architecture in Japan has designed high end homes, biennials and museums by using containers. Room heating is a primary concern of using container as housing material because steel boxes are good heat conductors. But by using passive or active cooling technology we provide insulation. Also it leaves a low carbon foot print too. As per right to shelter in U. P. Avas Vikas Parishad v. friends coop. Housing society limited, the right to shelter has been fundamental right to residence secured under article 19(1)e and the right to life guaranteed under article 21.The state has to provide facilities and opportunities to build houses to standardize life of poor people. Container housing is a new trend of green construction technology, also it is made up of a heavy, good engineering property steel which is corrosion free and has adequate life to use as shelter. Average cost of shipping container houses are ranging from 5-15 lakh INR. India uses about 4 to 5 % of total shipping containers used in the world. This paper covers the literature review, methodology adopted for research, challenges and opportunity to use container as housing material.

DOI: http://doi.org/10.5281/zenodo.20838604

Published by: vikaspatanker