Artificial Scenarios Research Articles (Page 1)

Thanks to advances in deep learning techniques, Human Pose Estimation (HPE) has achieved significant progress in natural scenarios. However, these models perform poorly in artificial scenarios such as painting and sculpture due to the domain gap, constraining the development of virtual reality and augmented reality. With the growth of model size, retraining the whole model on both natural and artificial data is computationally expensive and inefficient. Our research aims to bridge the domain gap between natural and artificial scenarios with efficient tuning strategies. Leveraging the potential of language models, we enhance the adaptability of traditional pose estimation models across diverse scenarios with a novel framework called VLPose. VLPose leverages the synergy between language and vision to extend the generalization and robustness of pose estimation models beyond the traditional domains. Our approach has demonstrated improvements of 2.26% and 3.74% on HumanArt and MSCOCO, respectively, compared to state-of-the-art tuning strategies.

Spinons are elementary excitations at the core of frustrated quantum magnets. Although it is well established that a pair of spinons can emerge from a magnon via deconfinement, controlled manipulation of individual spinons and direct observation of their deconfinement remain elusive. We propose an artificial gauge field scenario that enables the engineering of specific excited states in quantum spin models. This generates spatially localized individual spinons with high controllability. By applying time-dependent gauge fields, we realize adiabatic braiding of these spinons, as well as their dynamical evolution in a controllable manner. These results not only provide the first direct visualization of individual spinons localized in the bulk, but also point to new possibilities to simulate their confinement process. Finally, we demonstrate the feasibility of our scenario in Rydberg atoms, which suggests an experimentally viable direction-gauge-field engineering of correlated phenomena in excited states.

Intensive groundwater extraction and a severe 2021 drought have worsened land subsidence in Taiwan's Choshui Delta, highlighting the need for effective predictive modeling to guide mitigation. In this study, we develop a machine learning framework for subsidence analysis using electricity consumption data from pumping wells as a proxy for groundwater extraction. A long short-term memory (LSTM) neural network is trained to reconstruct missing subsidence records and forecast subsidence trends, while an artificial neural network links well electricity usage to groundwater level fluctuations. Using these tools, we identify groundwater-level decline from pumping as a key driver of subsidence. The LSTM model achieves high accuracy in reproducing historical subsidence and provides reliable predictions of subsidence behavior. Scenario simulations indicate that reducing groundwater pumping, simulated by lowering well electricity use, allows groundwater levels to recover and significantly slows the rate of land subsidence. To assess the effectiveness of pumping reduction strategies, two artificial scenarios were simulated. The average subsidence rate at the Xiutan Elementary School multi-layer compression monitoring well (MLCW) decreased from 2.23 cm/year (observed) to 1.94 cm/year in first scenario and 1.34 cm/year in second scenario, demonstrating the potential of groundwater control in mitigating land subsidence. These findings underscore the importance of integrating groundwater-use indicators into subsidence models and demonstrate that curtailing groundwater extraction can effectively mitigate land subsidence in vulnerable deltaic regions.

Bridges and viaducts play a fundamental role in the logistics infrastructure of a region. Early damage detection is essential for these assets′ managers to promote less invasive and lower-cost maintenance, and to extend their service life. This work presents an artificial intelligence (AI) based approach for detecting and locating damage in reinforced concrete bridges and viaducts. The methodology combines artificial damage scenarios obtained through a finite element model calibrated based on field tests and uses Artificial Neural Networks (ANN) to diagnose a case study viaduct’s condition. The model adjustment was achieved using field test data of the structure’s dynamic characteristics, and by adjusting the material properties of its structural elements and the boundary conditions. Multiple damage scenarios were created in key structural elements with different severity levels. Two neural networks were developed to diagnose the structure’s health: the first detects and locates the damage, and the second estimates its severity. The proposed algorithms exhibited the necessary accuracy for an accurate diagnosis. The proposed methodology can be used as an auxiliary tool for inspection and maintenance planning of the studied asset.

Power companies are increasingly under pressure to reconcile conventional hydrocarbon business with integrating renewables and staying profitable and within environmental goals. Digital twin technology is a game-changer in this respect, as it develops complex computational models of physical infrastructure that exchange data in real-time with field equipment. These systems integrate sophisticated computer vision for asset tracking, predictive analytics to anticipate equipment failure ahead of traditional means, and multi-objective optimization platforms aligning economic returns with sustainability targets. The architectural sophistication required to integrate offshore platform, refinery, wind farm, and solar installation data necessitates forward-looking microservice design and cloud infrastructure. Artificial intelligence integrated into these systems analyzes enormous volumes of data, uncovering latent relationships among weather conditions, equipment performance, and market conditions. This facilitates strategic, proactive maintenance plans and dispatch scheduling optimization over varied asset bases. Hybrid operating schemes showcase excellent synergies as solar thermal networks augment oil recovery processes while in-place pipeline infrastructure is readied for transporting hydrogen. Financial model innovation through generative AI produces artificial scenarios that challenge portfolio robustness to extreme market environments, while new hedging tools address weather-risked generation. The convergence of computational intelligence with physical infrastructure makes energy transformation from a burden into an opportunity, demonstrating that environmental responsibility and shareholder value proceed hand-in-hand through technological innovation.

Quantum machine learning has the potential to computationally outperform classical machine learning, but it is not yet clear whether it will actually be valuable for practical problems. While some artificial scenarios have shown that certain quantum machine learning techniques may be advantageous compared to their classical counterpart, evidence does not yet suggest that quantum machine learning has surpassed conventional approaches in dealing with standard classical datasets, such as the MNIST dataset. In contrast, dealing with quantum data, such as quantum states or circuits, may be the task where we can benefit from quantum methods. Therefore, it is important to develop practically meaningful quantum datasets for which we expect quantum methods to be superior. In this paper, we propose a machine learning task that is likely to soon arise in the real world: clustering and classification of quantum circuits. We provide a dataset of quantum circuits optimized by the variational quantum eigensolver. We utilized six common types of Hamiltonians in condensed matter physics, with a range of 4–20 qubits, and applied ten different ansatz with varying depths (ranging from 3 to 32) to generate a quantum circuit dataset of six distinct classes, each containing 300 samples. We show that this dataset can be easily learned using quantum methods. In particular, we demonstrate a successful classification of our dataset using real four-qubit devices available through IBMQ. By providing a setting and an elementary dataset where quantum machine learning is expected to be beneficial, we hope to encourage and ease the advancement of the field.

Virtual reality is part of the set of haptic technologies that allow the user to interact with an artificial scenario, different from the real world but very similar to it. Within the field of education, it promotes active and dynamic learning with immersive experiences. The goal of this study was to assess the usability of an object in VR format for oscilloscope instruction. We created a virtual physics lab environment in version 1.0. The methodology has a quantitative, descriptive approach. Brooke (1996) designed the System Usability Scale (SUS), which collected data from experts in both technology and physics, totaling 42 participants. We obtained Cronbach's alpha and McDonald’s omega indexes with values of 0.924 and 0.901, respectively. The results reveal that the virtual reality (VR) resource received an assessment of 72.58, corresponding to a percentile between 65 and 69, across 500 previous studies by various authors on the SUS. This indicates that the created object has a good acceptance range, being valued as very good and suitable for use in teaching.

Abstract For the field of telecommunications, the artificial intelligence (AI) has revolutionized the way of networking, due to improved scalability, reliability and efficiency networks. This research provides an insight into the role of artificial intelligence in telecom companies concentrating on its effectiveness on cost reduction, service quality, and network performance analysis. As traditional network growth they are seeing in customer demand for faster issue resolution and better management solutions. Artificial intelligence is one of the key enablers with the help of lowering the cost of operation and decreased energy usage while reducing the reliance on redundant manual actions. Involving objectifying the incident resolution in Bizagi by using simulations of two types of scenarios including artificial intelligence and scenario excluding the artificial intelligence. These scenarios were assessed both on cost, throughput and resource usage. Respondent level data were also collected from a survey that was administered to staff, supported the findings of the case study, data were analysed using Pearson and Tests of the equivalence of the proportions were performed via Chi-square tests in IBM SPSS Statistics. The simulation results reveal that AI based schemes improve the network resource consumption and we also find that energy consumption. speed up customer response times and improve demand management at peak times. Moreover, the incorporation of AI helps telecom companies to foreknow the issues in the service to effectively prevent the future failures. AI adopting plays important in telecom. This paper draws attention to the transformative power of artificial intelligence in sensing achiever in the engineering. Further, a questionnaire that were completed by with the company’s employees helped support the case study results whose analysis was done with Pearson and Chi-square analyses in IBM SPSS Statistics. The findings show that AI- enabled methods maximize network resources and maximize the inflation function. speedup customer reply times but also improve demand management in peak time. Moreover, the The embedded AI integration allows telecom operators to predict and prevent potential failures and hence Reduce the both downtime and downtime. reducing disruptions. This paper emphasizes the important power of artificial intelligence to process telecommunication optimization, which may provide useful for the companies who wish to increase functioning effectiveness and enhance customer satisfaction.

The Weber location problem is widely used in several artificial intelligence scenarios. However, the gradient of the objective does not exist at a considerable set of singular points. Recently, a de-singularity subgradient method has been proposed to fix this problem, but it can only handle the q-th-powered l_2-norm case (1

‘Wicked’ problems are among today's most complex and pressing issues. Examples of wicked problems are climate change and sustainability, as well as other problems that have critical implications for a wide range of stakeholders, including organizations and their leaders. Despite a growing body of work on wicked problems, solutions remain elusive, underscoring a need for new and innovative approaches. Recent technological advances are creating new opportunities to ‘untangle’ wicked problems by combining the use of artificial intelligence with scenarios. The purpose of this conceptual paper is therefore to present a framework for dealing more effectively with wicked problems by strengthening the utility of our responses to them. The first task is to define and explore the concept of a ‘wicked’ problem in order to identify suitable analytic entry points. This results in a proposed typology of wicked problems. Next, and by drawing from the literature, a set of relevant AI capabilities is synthesized. Their applicability to wicked problems is discussed and then operationalized by using the sequential steps in the risk management process framework. Using climate change as a focal issue, it is demonstrated that the combined use of AI and scenarios is synergistic, leaving us better prepared to tackle the challenges of wicked problems. The paper concludes with cautionary words, including about risks inherent in the use of AI. These have particular significance for leaders and the roles they should play.

The look-and-step behavior of biped robots requires quickly extracting planar regions and obstacles with limited computing resources. To this end, this paper proposes an efficient method representing the environment as a hybrid of feasible planar regions and a heightmap. The feasible planar regions are used for footstep planning, preventing the body from hitting obstacles, and the heightmap is used to calculate foot trajectory to avoid foot collision during the swing process. The planar regions are efficiently extracted by leveraging the organized structure of points for nearest neighbor searches. To ensure safe locomotion, these extracted planar regions exclude areas that could cause the robot's body to collide with the environment. The proposed method completes this perception process in 0.16 s per frame using only a central processing unit, making it suitable for look-and-step behavior of biped robots. Experiments conducted in typical artificial scenarios with BHR-7P and BHR-8P demonstrate its efficiency and safety, validating its effectiveness for the look-and-step behavior of biped robots.

Abstract In this study, an efficient surrogate‐assisted grey wolf optimizer (GWO) is presented by combining Kriging‐based active learning to identify damages in jacketed platforms based on modal analysis. The use of active learning in parallel with GWO significantly reduced the number of calls to the objective function and increased the accuracy of the algorithm's search in the problem space. The proposed approach was first evaluated on four benchmark problems, and its performance was validated against original GWO, particle swarm optimization (PSO), and genetic algorithm (GA) techniques. Then, by generating artificial damage scenarios on a real jacket platform in ABAQUS software, it was evaluated for the identification of damaged members. The results indicated high accuracy in estimation and an appropriate convergence rate in solving the high‐dimensional and complicated problem of damage detection of jacketed platforms. In such a way that the error rate of damage severity estimation in scenarios 1 and 2 was, on average, 3% and 5%, respectively. Meanwhile, the damage position was correctly estimated, and the call rate of the function was reduced by 50%. The efficiency of the proposed approach shows that it can be used for further works on the reliability‐based design of jacket structures.

The purpose of this paper is to classify and analyze the application scenarios of artificial intelligence in supply chain based on case studies of multinational companies. The paper first introduces the background and development of AI and discusses its current applications in supply chain management. Three main large-scale application areas are proposed, namely supply chain demand forecasting, risk management, and transportation operations planning. Three representative cases from Walmart, HP and UPS are reviewed based on the applied technologies, implemented processes, and advantages and disadvantages. As can be seen, this study contains meaningful recommendations to enhance the use of AI models in other industries and to adapt them to their characteristics. In conclusion, it can be said that AI has great potential to improve supply chain performance and resilience to adversity if conditions are taken into account in practical applications.

Artificial intelligence technologies are rapidly developing and having a major impact on the business world. Decision-making processes play an important role for the success of an organization. However, in today’s business world with its complexity and uncertainty, it becomes difficult to manage decision-making processes. At this point, creating future scenarios supported by artificial intelligence and working on different scenarios helps businesses to be more prepared for uncertainty. Artificial intelligence-supported scenarios can be utilized across various sectors and fields of work. AI enables businesses to analyze past data, predict trends, and consequently work on future scenarios to make more informed decisions. The significance of future scenarios lies in identifying risks and opportunities in advance, adapting to future changes, and being proactive in competition. By evaluating potential developments, shaping your business strategy, you can gain a competitive advantage and make more reliable decisions. Qualitative methods were employed in the research. Interviews were conducted with managers from 6 different professional groups (software, biomedical, public, construction, university, e-commerce). Data was collected and analyzed using semi-structured interview forms consisting of 4 questions. When the findings were evaluated, no concerns or negative expressions regarding the use of artificial intelligence were expressed. Except for public institutions, everyone has AI in their planning. Each sector believes it is important. No negative concerns were expressed. The prominent concepts in the findings are: Speed, big data, gaining competitive advantage, personalized customer experience, risk analysis, cost advantage, technology adaptation, optimization, accurate and fast situation detection, efficiency, etc. It is thought that the research will create significant awareness for businesses in the turbulent period of the 21st century, where uncertainties are greater than ever. Despite all the positive aspects, AI-supported decision-making processes also carry some risks. The most prominent risks include the applicability of AI-supported scenarios, security concerns, the existence of ill-trained AI models, ethical issues and data privacy.

Community Detection method based on Random walk and Multi objective Evolutionary algorithm in complex networks

Aim: To provide insights into the optimal use of virtual reality (VR) in nursing education by evaluating pre-registration nursing students' experiences in conducting holistic patient assessments while interacting with artificial intelligence (AI)-led patients. Specifically, this project evaluation compares the use of two different VR scenarios, one employing a menu-based interface and another using AI voice-controlled technology. Methods: Eleven pre-registration adult nursing students from two UK universities were selected through purposeful sampling to participate in the two VR simulations. Data collection and analysis: This included qualitative insights gathered from three focus group sessions, audio-recorded and thematically analysed to classify and describe students' experiences. Findings: Four key themes emerged: technological literacy, VR as a learning tool, the road of learning, and transition to independence. Advantages across both methods of VR-AI interaction and their particular challenges were identified and described for each key theme. Conclusion: VR with AI-led patient technology in pre-registration nursing education positively contributes to the curriculum by exposing students to problem-based learning situations and use of a multiplicity of skills in a safe environment. Although both methods are relevant for developing proficiencies around holistic patient assessment, there are advantages and limitations to each. Students perceived the voice-controlled technology as more intuitive with a more natural method of communication, whereas the menu-based interaction gave students more structure and guidance.

Artificial intelligence and machine learning methods build investment routes to balance models between private and public sources of financing. In this respect, they are of national importance for import substitution and technological sovereignty. Decision support systems build business development scenarios based on marked-up data. They reduce the risks of projects connected with import substitution and national technological sovereignty. Early integrated planning and balancing of developer and investor capabilities can help other venture and high-tech projects by balancing various sources of private and government financing. This article introduces a new development method of machine learning and artificial intelligence based on an ultraprecise neural network. The method automates the task of navigating technological projects using investment financing tools. It builds a continuous multi-agent investment route to reduce the risks of technological projects in terms of private and government investments. In fact, the method offers an algorithm that connects the fundraising stage, the type of project, and the type of funding source. The research objective was to strategize the development, implementation, and scaling of artificial intelligence methods and scenario multi-agent modeling to solve economic coordination tasks of raising public and private funds by personal investment routes and integrated investment routes. The authors rationalized the development, implementation, and scaling of personal and integrated investment routes, defined the development principles, and designed a checklist. They also developed a methodology for using artificial intelligence algorithms. The practical part featured a case of strategizing regional economic potentials in terms of raising additional funds by multi-agent modeling of financial and economic interaction of individual investment projects and integrated investment projects. The authors assessed the long-term multiplicative effect of investment projects on sectoral and intersectoral cooperation, which increases the regional investment attractiveness. The study relied on the theory of strategy and methodology of strategizing developed by Professor Vladimir L. Kvint.

Abstract Corrosion poses a significant threat to the longevity of steel bridges, impacting overall structural integrity. To effectively assess the structural condition of corroded steel bridges, conventional methods rely on visual inspections or single point measurements. To enhance and modernize this approach, this study introduces a novel framework integrating laser scanning data, computational models, and convolutional neural networks (CNNs). The CNN models are trained on a data set consisting of more than 1400 artificial corrosion scenarios generated by parameterizing real scan data from naturally corroded girders. This innovative method predicts the residual capacity and failure mode of corroded beam ends, achieving a low error rate of up to 3.3%. Unlike established evaluation procedures, the proposed evaluation framework directly utilizes post‐processed laser scanner output, eliminating the need for feature extraction and calculations.

Abstract Although cross-validation (CV) is a standard technique in machine learning and data science, its efficacy remains largely unexplored in ranking environments. When evaluating the significance of differences, cross-validation is typically coupled with statistical testing, such as the Dietterich, Alpaydin, or Wilcoxon test. In this paper, we evaluate the power and false positive error rate of the Dietterich, Alpaydin, and Wilcoxon statistical tests combined with cross-validation each operating with folds ranging from 5 to 10, resulting in a total of 18 variants. Our testing setup utilizes a ranking framework, similar to the Sum of Ranking Differences (SRD) statistical procedure: we assume the existence of a reference ranking, and distances are measured in $$L_1$$ L 1 -norm. We test the methods under artificial scenarios as well as on real data borrowed from sports and chemistry. The choice of the optimal CV test method depends on preferences related to the minimization of errors in type I and II cases, the size of the input, and anticipated patterns in the data. Among the investigated input sizes, the Wilcoxon method with eight folds proved to be the most effective, although its performance in type I situations is subpar. While the Dietterich and Alpaydin methods excel in type I situations, they perform poorly in type II scenarios. The inadequate performances of these tests raises questions about their efficacy outside of ranking environments too.

Age-related changes in time perception: Effects of immersive virtual reality and spatial location of stimuli