Authors: Aibek Tursunov

Abstract: The rapid evolution of artificial intelligence (AI) has transformed the way organizations approach customer relationship management, service delivery, and workflow automation. Salesforce, as a leader in enterprise software, has pioneered the integration of agentic architectures and hybrid AI-human collaboration models through platforms like Agentforce. This article explores the foundational principles of agentic architecture, its implementation within Salesforce, and the transformative impact of hybrid AI-human collaboration on business operations. Agentic architecture refers to the design of intelligent systems capable of sensing environments, making decisions, and acting autonomously within predefined parameters. These architectures are not monolithic; they encompass a spectrum of models, including single-agent, multi-agent, vertical, and horizontal systems, each suited to different operational needs. Hybrid architectures, which blend the strengths of these models, are particularly relevant in modern business environments where flexibility, scalability, and adaptability are paramount. Salesforce’s Agentforce platform exemplifies this hybrid approach by enabling seamless integration between AI agents and human workers. The platform leverages large action models to automate repetitive tasks, analyze vast datasets, and generate actionable insights, all while empowering human employees to focus on high-value, judgment-driven activities. This collaboration is not about replacing human labor but augmenting it, fostering a workforce where machines and people work in tandem to achieve superior outcomes. The article delves into the technical underpinnings of agentic architectures, the role of frameworks and controls in ensuring ethical and effective AI deployment, and the practical benefits of hybrid collaboration for businesses. By examining real-world use cases, metrics for success, and the evolving landscape of AI governance, this article provides a comprehensive overview of how Salesforce is redefining the future of work through agentic AI and hybrid collaboration.

DOI:

 

Published by: vikaspatanker

Authors: Mr. Adnan Shafiq Mangaonkar, Ritesh Kumar Indrajit Sharma

Abstract: India's rapid adoption of biometric technology has positioned it as a global leader in AI-driven identity recognition systems, with over 1.3 billion citizens enrolled in the Aadhaar database. This research examines the evolving trends and emerging threats in biometric data usage across India's digital ecosystem through secondary data analysis. The study analyzes five comprehensive case studies spanning government identification systems, law enforcement surveillance, banking sector authentication, consumer mobile applications, and healthcare implementations. Key findings reveal a biometric market valued at INR 24,303.6 crores in 2024, growing at 12.18% CAGR, alongside concerning security vulnerabilities including 815 million healthcare records breached in 2023 and a 300% increase in biometric data breaches between 2020-2023. The research identifies critical gaps in privacy frameworks, discriminatory policing practices using 80% accuracy thresholds, and inadequate regulatory oversight of consumer applications. While biometric authentication has enhanced financial inclusion and service delivery efficiency, significant threats include deepfake attacks, algorithmic bias, and mass surveillance capabilities. The study recommends strengthened data protection laws, transparent AI governance frameworks, and enhanced user consent mechanisms to balance technological innovation with fundamental privacy rights.

DOI: http://doi.org/10.5281/zenodo.15813266

 

Published by: vikaspatanker

Authors: Assistant Professor Mr. Sabir Nasir Mujawar

Abstract: Micro, Small, and Medium Enterprises (MSMEs) have emerged as a powerful force driving credit growth in the Indian banking sector, overtaking traditional retail lending segments. This research investigates the shifting dynamics of credit disbursement between MSMEs and retail loans using secondary data from RBI reports and case insights from Economic Times articles (2025). With improved asset quality and government-backed credit guarantee schemes such as CGFMU and ECLGS, banks are increasingly investing in MSME lending. This paper explores how digital infrastructure (e.g., TReDS platforms), fintech collaborations, and monetary policies have catalyzed MSME credit expansion. Through a case study and trend analysis of credit disbursal data from FY 2022–25, we examine sectoral credit performance, NPA levels, and structural challenges. The study concludes with strategic recommendations to sustain credit quality and unlock further growth in India’s MSME sector.

DOI: https://doi.org/10.5281/zenodo.15813033

Published by: vikaspatanker

Authors: Ms. Ifrah Kampoo, Ms. Tejashree Khandekar, Dr. Jasbir Kaur, Assistant Professor

Abstract: – In healthcare demand of giving a rapid solutions led in digital health services. The chatbot employs natural language processing to interpret user inputs and leverages supervised machine algorithms. We use Natural Language Processing(NLP)extracting the structure symptoms from the text. ML algorithm use like Naive Bayes use by Probabilistic mode good for text classification fast works with small datasets.Another algorithm Decision Tree use to Rule-based predicate from symptom combinations easy to interpret and fast. Random Forest also use by Ensemble of decision trees used in more accurate, handles noisy data well. Support Vector Machine(SVM) that algorithm used as Bnary/multi-class classification of symptoms high accuracy, effective with high-dimensional data. An associated symptom-disease dataset is used to train and validate the model. Technology is accelerating innovations in healthcare domain has increasing people living years.

DOI:

 

Published by: vikaspatanker

Authors: Ravi Prakash Jaiswal, Manish Saraf, Vijendra Pratap Singh, Ambuj Kumar Misra

Abstract: Nitrogen (N) deficiency remains a major constraint on cereal productivity because it reduces chlorophyll formation, canopy photosynthesis, and grain filling, while blanket fertilizer practices often fail to match within-field variability and reduce nitrogen use efficiency (NUE) (Govindasamy et al., 2023). Although destructive sampling and laboratory diagnostics are accurate, they are slow and difficult to scale for timely, spatially targeted decisions in real farms (Fu et al., 2021). This study frames N deficiency detection as a cross-domain transfer learning problem and develops a cross-crop machine learning framework for wheat, maize, and rice using RGB imagery under field conditions. We harmonized and profiled three public datasets (wheat: 1,381 leaf images; maize: 1,200 canopy/plot images; rice: 1,500 leaf images with Leaf Color Chart-based labeling), applied standardized preprocessing, and trained baseline CNN and fine-tuned ResNet models with fixed random seeds and identical train/validation/test splits for reproducibility. Performance was evaluated under three scenarios: within-crop testing, direct cross-crop transfer without retraining, and domain adaptation using unlabeled target data. Four adaptation methods were benchmarked: CORAL, MMD, AdaBN, and Domain-Adversarial Neural Networks (DANN) (Ganin et al., 2016; Gretton et al., 2012; Li et al., 2016; Sun & Saenko, 2016). Baseline cross-crop transfer showed substantial generalization gaps (≈25–35 percentage points), with accuracy ranging from 47.6% to 56.2% across crop pairs, confirming severe domain shift (Fu et al., 2021; Pan & Yang, 2010). Domain adaptation improved average cross-crop accuracy from 51.7% (baseline) to 58.3% (AdaBN), 60.1% (CORAL), 64.6% (MMD), and 73.2% (DANN), with DANN delivering up to ~19% absolute improvement and the most consistent gains under challenging transfers (Ganin et al., 2016). Overall, results indicate that adversarial domain adaptation can substantially reduce cross-crop failure modes and supports more scalable nitrogen monitoring with reduced dependence on crop-specific labels, while practical deployment should include agronomic guardrails and uncertainty-aware decision rules for safe in-season recommendations (Fu et al., 2021).

DOI: https://doi.org/10.5281/zenodo.18514672

Published by: vikaspatanker

Authors: Ndongkeh Nelson Maineh, Edith Bate Etakah, Edna Buhnyuy Visiy, Mbeck Prosper Wanlo

Abstract: Plastic and diaper waste are major pollution problems worldwide. Approximately 72% of global plastic and diaper wastes end up in landfills, exacerbating environmental degradation, highlighting the urgent need for valorization strategies. This study investigated the potential use of Polyethylene Terephthalate (PET) plastics, diaper wastes and sand for the production of pavement blocks, with the goal of developing an environmentally sustainable method for repurposing these waste materials into valuable construction products. Four formulations of the paving blocks were produced and their mechanical and physical properties evaluated through various testing methods. For the four formulations, the plastic (binder) content was maintained at a constant 45% while the diaper (aggregate) content was being varied across formulations, replacing sand at percentages of 0%, 2.5%, 5% and 10% respectively. The results showed that the compressive strength of the blocks remained relatively constant across the first three formulations, with values of 10.23 MPa, 10.25 MPa, and 10.25 MPa, respectively, but dropped significantly in the fourth formulation (5.57 MPa). This indicated that a 10% replacement of sand by diapers in the fourth formulation is not advisable, as their compressive strength falls below the minimum of 8.5 MPa recommended by the SNI 03-0691-1996 standard. Moreover, the results also indicated that both flexural strength and abrasion resistance of the blocks declined as the diaper concentration increased, suggesting an optimal threshold concentration for incorporating waste diapers into the waste blocks. Also, the water absorption rate of the paving blocks increased with increasing diaper concentration with values of 0.34%, 0.34%, 0.92%, and 2.18%, respectively. However, all values were within the <20% limit for high quality blocks (ISS 1077-1970 Standard) suggesting that the blocks can withstand extreme environmental conditions, such as floods. The research demonstrates the potential to co-valorize waste plastics, diapers, and sand for the production of sustainable pavement blocks.

DOI: https://doi.org/10.5281/zenodo.15811965

Published by: vikaspatanker

Authors: MCA,M.Phil.,, Ms.S.Prathi

Abstract: – Multimodal sentiment analysis (MSA) integrates data from multiple sources, such as text, audio, and visual cues, to enhance the accuracy and interpretability of sentiment classification models. Traditional sentiment analysis predominantly relies on textual data, which can be limited in capturing non-verbal nuances like tone of voice or facial expressions. This paper explores the synergy between text, speech, and visual data in sentiment analysis tasks, addressing key challenges such as data alignment, feature extraction, and fusion techniques. We compare various fusion strategies, including early, late, and hybrid fusion, using state-of-the-art deep learning architectures such as convolutional neural networks (CNNs), recurrent neural networks (RNNs), and transformers. Experimental results demonstrate that multimodal approaches significantly outperform unimodal systems, providing higher accuracy and robustness in sentiment detection. We discuss the potential applications of multimodal sentiment analysis in fields such as social media monitoring, customer sentiment analysis, and healthcare. Finally, the paper outlines future research directions, emphasizing the need for more efficient fusion techniques and the incorporation of emerging models to advance multimodal sentiment analysis further.

DOI: https://doi.org/10.5281/zenodo.15811942

 

Published by: vikaspatanker

Authors: Indra Bhuwan Yadav, B.Tech (C.S.E),, Neeharika Sengar, Assistant Professor, SOET, Dr. Rajendra Singh, HOD, SOET

Abstract: With the surge in digital services, backend infrastructure must manage high-volume transactions across diverse platforms with minimal latency and high reliability. Synchronous systems often struggle under concurrent loads, leading to performance bottlenecks. To address these limitations, this research introduces TRANSACTINET, a backend framework developed using the Tornado Python library. Designed with an event-driven model, the system supports asynchronous processing, modular architecture, and robust security protocols. This paper outlines its design, deployment strategy, and performance evaluation across real-world applications such as banking and agent-based commerce.

DOI:

 

 

Published by: vikaspatanker

Authors: Ms. Komala R, Shreya Sankannavar,

Abstract: The farming industry is a mainstay of the world economy, with potato cultivation contributing immensely to food security. Despite this, potato plants are very prone to several diseases like Earl y Blight, Late Blight, and bacterial infections, causing them to experience tremendous losses in yields. Conventional methods of disease detection involve the use of manual checking, which is time-consuming, labor-intensive, and inaccurate because of human error. To overcome such challenges, the project suggests a mac hine learning-based automatic potato disease detection system. The suggested system applies image processing and deep learning models to identify and classify diseases from leaf images with high accuracy. A dataset of healthy and diseased potato leaf images is preprocessed and utilized for training a convolutional neural network (CNN) model. The model is trained to classify different diseases and healthy leaves by learning from visual attributes. After training, the model detects diseases with high accuracy in real- time, allowing timely intervention and minimizing crop loss. This framework can help farmers and agricultural professionals keep track of crop health more effectively, increasing productivity and encouraging sustainable agriculture. The project indicates the use of machine learning in precision farming and how it can revolutionize conventio nal farming practices.

DOI: http://doi.org/

 

 

Published by: vikaspatanker

Authors: Professor Dr. Satya Singh, Ratnesh Kumar Sharma

Abstract: COVID-19 harmed the lives of people in every region of the world. It has been established that, in addition to the physical symptoms, it significantly influences the patient’s mental health. Depression has been identified as one of the most widespread disorders that can hasten a person’s mortality at an early age. This is one of the conditions that has been singled out for this distinction. The trajectory of life for millions of people has been altered as a result of this illness. We conducted a survey that consisted of 21 questions based on the Hamilton instrument and the advice of a psychiatrist. This was done so that we could continue forward with the inquiry into the identification of depression in individuals. After the data were compiled and analysed, it became clear that people younger than 45 years of age had a higher risk of suffering from depression when compared to those older than 45 years of age. This is because most people at this age are concerned about getting married or schooling their children. On the other side, research has revealed that those whose ages fall between 18 and 25 are also at an increased risk of suffering from depression. This is likely because, at this stage in their lives, these individuals are more conscious of the potential outcomes of their lives. Based on all of the replies received, the findings of the survey were put through several different machine learning algorithms, including Decision Tree, KNN, and Naive Bayes. These algorithms were used to analyse the results. Further investigation is being done into how these two techniques are similar to and different from one another. According to the findings of the research, KNN has produced better results than other approaches in terms of accuracy, whereas decision trees have produced better results in terms of the amount of time needed to detect depression in a person. In conclusion, to overcome the traditional approach to a depression diagnosis, which is made up of affirmative questions and constant feedback from individuals, a model that is based on machine learning is offered as a potential alternative.

DOI: https://doi.org/10.5281/zenodo.15804697

 

Published by: vikaspatanker