ResearchBerg
Not a member yet
161 research outputs found
Sort by
Contextual Understanding in Neural Dialog Systems: the Integration of External Knowledge Graphs for Generating Coherent and Knowledge-rich Conversations
Get PDFThe integration of external knowledge graphs has emerged as a powerful approach to enrich conversational AI systems with coherent and knowledge-rich conversations. This paper provides an overview of the integration process and highlights its benefits. Knowledge graphs serve as structured representations of information, capturing the relationships between entities through nodes and edges. They offer an organized and efficient means of representing factual knowledge. External knowledge graphs, such as DBpedia, Wikidata, Freebase, and Google\u27s Knowledge Graph, are pre-existing repositories that encompass a wide range of information across various domains. These knowledge graphs are compiled by aggregating data from diverse sources, including online encyclopedias, databases, and structured repositories. To integrate an external knowledge graph into a conversational AI system, a connection needs to be established between the system and the knowledge graph. This can be achieved through APIs or by importing a copy of the knowledge graph into the AI system\u27s internal storage. Once integrated, the conversational AI system can query the knowledge graph to retrieve relevant information when a user poses a question or makes a statement. When analyzing user inputs, the conversational AI system identifies entities or concepts that require additional knowledge. It then formulates queries to retrieve relevant information from the integrated knowledge graph. These queries may involve searching for specific entities, retrieving related entities, or accessing properties and attributes associated with the entities. The obtained information is used to generate coherent and knowledge-rich responses. By integrating external knowledge graphs, conversational AI systems can augment their internal knowledge base and provide more accurate and up-to-date responses. The retrieved information allows the system to extract relevant facts, provide detailed explanations, or offer additional context. This integration empowers AI systems to deliver comprehensive and insightful responses that enhance user experience. As external knowledge graphs are regularly updated with new information and improvements, conversational AI systems should ensure their integrated knowledge graphs remain current. This can be achieved through periodic updates, either by synchronizing the system\u27s internal representation with the external knowledge graph or by querying the external knowledge graph in real-time
An Empirical Evaluation Framework for Autonomous Vacuum Cleaners in Industrial and Commercial Settings: A Multi-Metric Approach
Get PDFDespite advancements in cleaning automation, there is a noticeable gap in standardized evaluation methods for autonomous vacuum cleaners in industrial and commercial settings. Existing assessments often lack a unified approach, focusing narrowly on either technical capabilities or financial aspects, without integrating both perspectives. This research presents a framework for the evaluation of autonomous vacuum cleaners in industrial and commercial settings, focusing on eight key metrics. These metrics are designed to provide a unified empirical perspective of the vacuum cleaners\u27 performance, operational efficiency, cost, productivity, durability, safety, return on investment, and adaptability. The proposed framework starts with an analysis of cleaning efficiency, examining both the area covered by the cleaners and the quality of cleaning. Advanced image processing techniques are suggested for mapping the area coverage, tailored to different vacuum designs. For assessing cleaning quality, the proposal highlights the potential integration of real-time dirt detection technologies, such as gravimetric sampling and light sensors, to dynamically adapt to varying dirt concentrations and types. Operational efficiency part encompasses the assessment of battery life, charge time, and operational downtime. It advocates for a dual approach of empirical testing and analytical modeling to measure battery life and charge time accurately. The evaluation of operational downtime incorporates tracking of maintenance, charging periods, and other non-operational activities, complemented by predictive modeling for efficient future planning. The financial aspect of the proposed framework encompassed under cost metrics, considers the initial investment, operational and maintenance costs, and potential labor cost savings. This study argues that these cost analysis aids in understanding the long-term financial implications of adopting autonomous vacuum cleaners. Productivity metrics focus on the cleaning speed and the level of autonomy of the vacuum cleaners. Cleaning speed is evaluated using formulas that take into account various environmental factors, while the autonomy level is determined using Sheridan\u27s Levels of Autonomy, which reflects the vacuum\u27s operational independence and its impact on human productivity. Durability, reliability, safety, and compliance are key for vacuum cleaners, evaluated through metrics like Mean Time Between Failures, Mean Time To Repair, Service Life, safety incidents, and adherence to standards and regulations. Lastly, the suggested framework evaluates the vacuum\u27s flexibility and adaptability in different environments, such as various floor types and conditions, highlighting the importance of versatility in autonomous cleaning solutions.
Article history: Received: 01/December /2022; Available online: 07/ February/2023; This work is licensed under a Creative Commons International License
Career Advancement Barriers Faced by LGBTQ Employees: An Exploration of Discrimination, Bias, and Inclusion in the Workplace
Get PDFBackground: LGBTQ individuals have historically faced discrimination and stigma in the workplace, and this can impact their career advancement opportunities. Despite increasing awareness and legal protections, it is still unclear how pervasive these barriers are and how they impact LGBTQ employees\u27 advancement opportunities.
Findings: Our study aimed to explore career advancement barriers faced by LGBTQ employees. We conducted a survey of 500 LGBTQ employees across various industries in the United States. Our findings show that workplace discrimination, lack of mentorship and sponsorship, bias in hiring and promotion, fear of being out, and lack of inclusive policies and benefits are significant barriers to career advancement for LGBTQ employees. Workplace discrimination was found to be the most significant barrier, with 75% of respondents reporting that they had experienced some form of discrimination in the workplace due to their LGBTQ identity.
Recommendations: To address these barriers, we recommend that workplaces take several steps. Firstly, workplaces must implement policies that protect against discrimination, including sexual orientation and gender identity. This can be achieved by implementing diversity and inclusion training for all employees and holding managers and employees accountable for discriminatory behavior. Secondly, offering mentorship and sponsorship programs to LGBTQ employees is crucial to their career advancement. Thirdly, workplaces must ensure that LGBTQ employees have equal opportunities for hiring and promotion, and this can be achieved by implementing objective criteria for evaluation and decision-making. Fourthly, creating a supportive and inclusive environment where LGBTQ employees can feel safe and supported to be out is essential. Finally, workplaces should offer benefits that are inclusive of LGBTQ employees, such as healthcare coverage for gender-affirming procedures and parental leave for same-sex couples.
Conclusion: Our findings highlight the significant barriers that LGBTQ employees face in career advancement due to discrimination, bias, and lack of support. Our recommendations provide a roadmap for workplaces to create a more inclusive and supportive environment for LGBTQ employees, which benefits both the employees and the organization as a whole
AI and the Future of Cognitive Decision-Making in HR
Get PDFThis study examines the impact of Artificial Intelligence (AI) on managerial cognition in the field of Human Resource Management (HRM). The research explores how AI systems enhance decision-making by reducing cognitive load, providing data-driven insights, and improving problem-solving capabilities in complex HR tasks such as recruitment, workforce planning, and performance management. A key focus of the study is on balancing AI-generated insights with human intuition, especially in areas where qualitative judgment, such as assessing cultural fit and leadership potential, remains essential. The study also highlights the cognitive adjustments HR managers must make to collaborate effectively with AI, emphasizing the development of data literacy and critical engagement with AI outputs. Additionally, the research addresses the risks of over-reliance on AI, including automation bias and the perpetuation of cognitive biases, and suggests strategies for mitigating these risks through active human oversight. Ethical considerations, including transparency, fairness, and accountability, are explored, and the study advocates for the integration of ethical frameworks into AI systems to ensure responsible and unbiased decision-making. This study contributes to the growing understanding of how AI can enhance, rather than replace, managerial cognition in HRM, fostering more efficient and ethical decision-making processes
Comparative Study of Cloud Encryption Algorithms and Their Role in Securing Data Across Public, Private, and Hybrid Cloud Environments
Get PDFThis paper presents a comparative analysis of cloud encryption algorithms and their role in securing data across public, private, and hybrid cloud environments. As cloud adoption grows, the need for robust encryption techniques to protect sensitive data from unauthorized access, breaches, and internal threats becomes increasingly critical. The study focuses on two primary encryption approaches: symmetric encryption, exemplified by Advanced Encryption Standard (AES), and asymmetric encryption, represented by RSA. It examines the strengths and weaknesses of these algorithms in different cloud models, with special attention to key management, computational efficiency, scalability, and security. Public clouds, characterized by multi-tenancy and external threats, often employ AES for its performance, while private clouds, with their controlled environments, may use a mix of AES and RSA to secure internal communications and data storage. Hybrid clouds pose additional challenges, as data needs to be encrypted seamlessly across public and private environments. This paper also highlights future directions, including the potential of emerging technologies such as homomorphic encryption and quantum-resistant algorithms to address existing challenges, particularly around key management and computational overhead. The findings aim to guide organizations in selecting encryption strategies tailored to their specific cloud environments, balancing security and performance requirements
Mitigating Cross-Site Request Forgery (CSRF) Attacks Using Reinforcement Learning and Predictive Analytics
Get PDFCross-Site Request Forgery (CSRF) attacks pose a significant threat to web application security, allowing attackers to perform unauthorized actions on behalf of authenticated users. Traditional CSRF mitigation techniques, such as using secure tokens and validating request origins, have limitations in adapting to attack patterns and optimizing security policies. This research explores the application of reinforcement learning (RL) and predictive analytics to enhance CSRF mitigation strategies. We propose several RL-based approaches, including CSRF token generation, CSRF detection, request validation, user behavior analysis, and security policy optimization. In these approaches, RL agents are trained to generate secure tokens, detect CSRF attacks, validate request authenticity, model user behavior, and optimize security policies based on observed attack patterns and system performance. The agents learn through simulated attack scenarios, real-world web traffic data, and continuous feedback, adapting to new CSRF techniques and balancing security effectiveness with user experience. Additionally, we investigate predictive analytics techniques for CSRF mitigation, such as anomaly detection, risk scoring, user behavior analysis, predictive token generation, and adaptive security policies. These techniques leverage machine learning algorithms to identify anomalous requests, assign risk scores, classify user behavior, generate secure tokens, and dynamically adjust security measures based on predicted risk levels. The research demonstrates the applications of RL and predictive analytics in enhancing CSRF mitigation strategies. These approaches offer promising solutions to strengthen web application security by proactively detecting and preventing CSRF attacks, adapting to attack patterns, and optimizing security policies. Further research is needed to validate the practicality and scalability of these techniques in real-world deployments and to integrate them with existing CSRF mitigation best practices. This research contributes to the field of web application security by introducing innovative approaches that leverage RL and predictive analytics to mitigate CSRF attacks. The proposed techniques may significantly improve the resilience of web applications against CSRF threats
Threat Mitigation in Containerized Environments
Get PDFContainerized environments have revolutionized application development and deployment, offering unmatched flexibility, scalability, and consistency across diverse infrastructures. However, these environments also introduce unique security challenges, stemming from the shared nature of resources, the immutability of container images, and the complexities of container orchestration platforms like Docker and Kubernetes. This paper provides an in-depth exploration of the threat landscape within containerized environments, focusing on key areas such as container escape, image vulnerabilities, network attacks, and supply chain risks. We also discuss robust mitigation strategies, including the use of hardened images, network segmentation, and container-specific security tools. The analysis culminates in a set of best practices aimed at securing containerized environments against evolving threats
Pharmacovigilance Monitoring of Herbal and Traditional Medicine Products: A Survey of Challenges and Solutions
Get PDFThis study explores the challenges and solutions in pharmacovigilance monitoring of herbal and traditional medicine products. The findings indicate that monitoring the safety and efficacy of these products is a complex task that requires a multidisciplinary approach. The lack of standardization in herbal and traditional medicine products poses a significant challenge in assessing their safety and efficacy. The absence of standardized production, quality control, and labeling methods can lead to variations in potency, purity, and composition. However, the study found that developing standardized methods for production, quality control, and labeling can enhance the safety and efficacy of herbal and traditional medicine products. The limited scientific evidence supporting the safety and efficacy of these products is another significant challenge. The study recommends encouraging and supporting rigorous scientific studies, such as randomized controlled trials, to provide evidence-based safety and efficacy data. The study also identified cultural and linguistic barriers as a challenge in monitoring the safety of herbal and traditional medicine products. The diverse cultural and linguistic backgrounds of users make it difficult to identify and report adverse reactions. To address this, the study recommends increasing public awareness and education about the safe use of herbal and traditional medicine products, including the potential risks and benefits. The lack of regulation of herbal and traditional medicine products is another significant challenge. The study recommends improving the regulation of herbal and traditional medicine products to ensure their safety, efficacy, and quality. The study also identified difficulty in identifying the active ingredient in some herbal and traditional medicine products as a challenge. Some products contain multiple ingredients, making it difficult to identify the active ingredient responsible for adverse effects. The study recommends developing appropriate methods for identifying the active ingredients in herbal and traditional medicine products. The study found that adverse events associated with herbal and traditional medicine products may be underreported, as many consumers may not recognize or report adverse effects. To address this, the study recommends encouraging and facilitating the reporting of adverse effects associated with herbal and traditional medicine products by healthcare professionals and consumers. We recommend fostering collaboration between healthcare professionals, regulatory agencies, manufacturers, and consumers to ensure the safety and efficacy of herbal and traditional medicine products. This would enable the development and implementation of appropriate pharmacovigilance systems for these products, ensuring continuous monitoring of their safety and efficacy. The study underscores the need for a multidisciplinary approach involving healthcare professionals, regulatory agencies, manufacturers, and consumers to address the challenges in monitoring the safety and efficacy of herbal and traditional medicine products
Opportunities and Challenges of Cloud Computing in Developing Countries
Get PDFCloud computing presents developing nations with a multitude of options, including enhanced access to technologies and services, higher productivity and cost savings, as well as the possibility of increased economic development and the creation of new jobs. It is possible for companies to have access to enterprise-grade software and tools if they have improved access to technology and services. This may assist the organizations increase their productivity and efficiency. The delivery of essential public services like healthcare, education, and social welfare may be improved via increased efficiency and cost reductions, which can also assist bring down overall prices. The formation of new industries and enterprises, in addition to the expansion and improvement of existing ones, may help to boost economic growth and the production of new jobs. However, cloud computing presents developing nations with a number of obstacles, including inadequate internet infrastructure, a lack of available technological skills, and concerns over the privacy and security of stored data. Because of limited internet infrastructure, accessing cloud services, transferring data, and implementing security measures might be challenging, which can restrict the usage of cloud-based services. When it comes to setting up and maintaining cloud-based systems, as well as selecting the appropriate cloud service providers and solutions, it may be challenging for businesses that lack the technical experience. Concerns over data security and privacy may be a significant obstacle to the widespread adoption of cloud computing. This is because businesses may lack confidence in the safety of their data if it is housed on servers located in other countries
The Role Artificial Intelligence in Modern Banking: An Exploration of AI-Driven Approaches for Enhanced Fraud Prevention, Risk Management, and Regulatory Compliance
Get PDFBanking fraud prevention and risk management are paramount in the modern financial landscape, and the integration of Artificial Intelligence (AI) offers a promising avenue for advancements in these areas. This research delves into the multifaceted applications of AI in detecting, preventing, and managing fraudulent activities within the banking sector. Traditional fraud detection systems, predominantly rule-based, often fall short in real-time detection capabilities. In contrast, AI can swiftly analyze extensive transactional data, pinpointing anomalies and potentially fraudulent activities as they transpire. One of the standout methodologies includes the use of deep learning, particularly neural networks, which, when trained on historical fraud data, can discern intricate patterns and predict fraudulent transactions with remarkable precision. Furthermore, the enhancement of Know Your Customer (KYC) processes is achievable through Natural Language Processing (NLP), where AI scrutinizes textual data from various sources, ensuring customer authenticity. Graph analytics offers a unique perspective by visualizing transactional relationships, potentially highlighting suspicious activities such as rapid fund transfers indicative of money laundering. Predictive analytics, transcending traditional credit scoring methods, incorporates a diverse data set, offering a more comprehensive insight into a customer\u27s creditworthiness. The research also underscores the importance of user-friendly interfaces like AI-powered chatbots for immediate reporting of suspicious activities and the integration of advanced biometric verifications, including facial and voice recognition. Geospatial analysis and behavioral biometrics further bolster security by analyzing transaction locations and user interaction patterns, respectively. A significant advantage of AI lies in its adaptability. Self-learning systems ensure that as fraudulent tactics evolve, the AI mechanisms remain updated, maintaining their efficacy. This adaptability extends to phishing detection, IoT integration, and cross-channel analysis, providing a comprehensive defense against multifaceted fraudulent attempts. Moreover, AI\u27s capability to simulate economic scenarios aids in proactive risk management, while its ability to ensure regulatory compliance automates and streamlines a traditionally cumbersome process