Search

This study aimed to investigate the barriers to responsible interaction between the global employer, technology provider, and trade union to realize the postulate of ethical artificial intelligence in the algorithmic management process. To overcome the barriers, this study offers a pedagogical explanation based on a universal schema for shaping a machine-learning model

The relevance of the research is conditioned by the fact that artificial intelligence technologies are rapidly developing and have enormous potential, which can be successfully used for the benefit of humanity, but at the same time they pose many new challenges in conditions of social and ethical and legal uncertainty and pose a number of ethical questions regarding the future way of life of human society and ways, which will be followed by its further development. The article emphasizes that artificial intelligence has the potential to change the future of humanity for the better, however, artificial intelligent systems are inherently not neutral and are characterized by inherent bias, which is due to the initial data used in their “training”. Due to the magnitude of the social consequences of artificial intelligence technologies, many countries are now concerned about the ethical aspects of its use. In order to identify possible scenarios and use the potential of artificial intelligence to realize development opportunities while maintaining risk control, it is important to develop a more comprehensive understanding of the social changes caused by the increasingly expanding use of intelligent systems. It is concluded that the problems of using artificial intelligence must be considered through the prism of analyzing the social essence of a person, modern socio-cultural reality, humanistic goals and values of modern society. Any achievements in the field of artificial intelligence make sense only if they correspond to truly human values. To date, the absence of internationally approved ethical and legal norms and standards related to the application of developments and innovations in the field of artificial intelligence indicates the need for active and focused work of the Russian research community in this area.

This presentation aims to explore the intricate balance between harnessing the potential of AI for societal advancement and addressing the ethical implications of data utilization. Nowadays data centers where user information is stocked must follow a very specific set of laws. This empowers researchers in the field of AI to train their models without feeding them with decrypted user data. But who decides how to use the data? Currently most companies that gather information from their customers must follow a set of directives from the EU Commission. Therefore, the evolution of the artificial intelligence would not mean the invasion of personal privacy. In this paper we hope to increase the necessity of treating data as legal object to which will apply reinterpreted rules, according to the flow of AI development. We will focus more on the content of data and to the moment when the data might be stored and transferred because of its increasing value (economic value). Data is an asset who can be traded. Also, data can help to create a personality of the users (this is why the value of data can increase). Another question we will try to answer concerns the concept of „digital estate” and postmortem data. After a person dies what will happen with its digital form like digital money (bitcoin or assets won in a game), with the Facebook account, Google account etc.? Data from people will make artificial people who will survive the natural death?

The size of urban settlements is rapidly increasing worldwide. This sprawl triggers changes in land cover with the consumption of natural areas and affects ecosystems with important ecological, climate, and social transformations. Detecting, mapping, and monitoring the growth and spread of urban areas is therefore important for urban planning, risk analysis, human health, and biodiversity conservation. Satellite images have long been used to map human settlements. The availability of the Sentinel constellation (S2) allows the monitoring of urban sprawl over large areas (e.g., countries) and at high frequency (with possible monthly updates). This massive data stream allows the proposal of new types of urban products at a spatial resolution of 10 meters. In this context, we developed a fully automated and supervised processing chain (URBA-OPT) using open-source libraries and optimized for rapid calculation on high-performance computing clusters. This processing chain has been applied to all of France. The objective is to propose an annual product at 10 m spatial resolution available through the THEIA Data and Services Centre. Our results demonstrate the feasibility and accuracy of producing an annual urban settlement map using this methodology, providing a valuable tool for urban planning and environmental monitoring.

The integration of renewable energy into transportation systems is essential for reducing greenhouse gas emissions and achieving global sustainability goals. This review explores the challenges, innovations, and future directions of incorporating renewable energy sources such as solar, wind, and bioenergy into transportation infrastructures. Key challenges include the intermittency of renewable energy, the need for advancements in energy storage systems, and the regulatory and economic barriers hindering widespread adoption. Innovations such as electric vehicles (EVs), vehicle-to-grid (V2G) technologies, and smart grids are pivotal in enabling this transition. Furthermore, artificial intelligence (AI) and machine learning (ML) offer significant potential to optimize energy management, enhance efficiency, and facilitate the smooth integration of renewable energy with transportation systems. The review also discusses successful case studies from different regions and examines policy frameworks supporting renewable energy in transportation. Future directions point toward increased collaboration between industries, technological advancements, and supportive policies to create a more sustainable, resilient transportation sector. Ultimately, this review aims to provide a comprehensive understanding of how smart integration of renewable energy into transportation can drive a cleaner, more sustainable future.

Estimating the ‘State of Charge’ (SOC) is a complex endeavour. Data-driven techniques for SOC estimation tend to offer higher prediction accuracy compared to traditional methods. With the progression of Artificial Intelligence (AI), machine learning has found extensive applications across various fields such as infotainment, driver assistance systems, and autonomous vehicles. This paper categorizes the machine learning techniques utilized in Battery Management System (BMS) applications and employs a modern supervised neural network approach to predict SOC. Accurate SOC estimation is crucial to prevent battery failures in critical situations, such as during heavy traffic or when traveling with limited access to charging stations. Long Short-Term Memory (LSTM) networks are particularly adept at classifying, processing, and predicting based on time series data. These models are capable of capturing and retaining features over time, making them suitable for this study. The model's predicted SOC closely matches the true SOC, and the SOC prediction error remains nearly zero even with a large sample of input data.

The introduction of automated algorithmic processes (e.g. machine learning) in creative disciplines such as architecture and urban design has expanded the design space available for creativity and speculation. Contrary to previous algorithmic processes, machine learning models must be trained before they are deployed. The two processes (training and deployment) are separate and, crucially for this paper, the outcome of the training process is not a spatial object directly implementable but rather code. This marks a novelty in the history of the spatial design techniques which has been characterised by design instruments with stable properties determining the bounds of their implementation. Machine Learning models, on the other hand, are design instruments resulting from the training they undertake. In short, training a machine learning model has become an act of design. Beside spatial representation traditionally comprising of drawings, physical or CAD models, Machine Learning introduces an additional representational space: the vast, abstract, stochastic, multi-dimensional space of data, and their statistical correlations. This latter domain – broadly referred to as latent space – has received little attention by architects both in terms of conceptualising its technical organisation and speculating on its impact on design. However, the statistical operations structuring data in latent space offer glimpses of new types of spatial representations that challenge the existing creative processes in architectural and urban design. Such spatial representation can include non-human actors, give agency to a range of concerns that are normally excluded from urban design, expand the scales and temporalities amenable to design manipulation, and offer an abstract representation of spatial features based on statistical correlations rather than spatial proximity. The combined effect of these novelties that can elicit new types of organisation, both formally and programmatically. In order to foreground their potential, the paper will discuss the impact of ml models in conjunction with larger historical and theoretical questions underpinning spatial design. In so doing, the aim is not to abdicate a specificity of urban design and uncritically absorb computational technologies; rather, the creative process in design will provide a filter through which critically evaluate machine learning techniques. The paper tasks to conceptualise the potential of latent space design by framing it through the figure of the paradigm. Paradigms are defined by Thomas Kuhn as special members of a set which they both give rise to and make intelligible. Their ability to relate parts to parts not only resonates with the technical operations of ml models, but they also provide a conceptual space for designers to speculate different spatial organisation aided by algorithmic processes. Paradigms are not only helpful to conceptualise the use of ml models in urban design, they also suggest an approach to design that privileges perception over structure and curation over process. The creative process that emerges is one in ml models are speculative technical elements that can foreground relations between diverse datasets and engender an urbanism of relations rather than objects. The application of such algorithmic models to design will be supported by the research developed by students part of Research Cluster 14 part of the Master in Urban Design at The Bartlett School of Architecture in London.

This study evaluates deep learning (DL) models, particularly ResU-Net with attention mechanisms, for mapping landslides in Quebec, Canada, utilizing high-resolution digital elevation model (HRDEM) data and its seven derivatives (slope, aspect, hillshade, curvature, ruggedness, surface area ratio, and max difference from mean). Three scenarios were considered to assess the effectiveness of various features in landslide segmentation: training the model on all features, each feature individually, and on slope and hillshade. Model performance on individual features was significantly poor, while the model trained with hillshade and slope outperformed the model using all seven features, particularly in F1-score (improved by 8% for rotational landslides and 11% for retrogressive landslides) during validation. Furthermore, for the test dataset, model performance on all seven features was compared against slope and hillshade. As a result, for rotational landslides, slope and hillshade achieved F1-scores of 0.68 and 0.93 for rotational and retrogressive landslides, respectively, while the same metrics using all features were 0.61 and 0.83, respectively. This suggests hillshade and slope provide the most relevant information and reduce computational complexity. Overall, the findings enhance our understanding of HRDEM derivatives and emphasize the importance of feature selection in optimizing model performance and reducing computational complexity.

Portfolio optimization plays a crucial role in investment decision-making by balancing risk and return objectives. With the aim of improving portfolio performance, while enhancing risk management, this literature review explores traditional and artificial intelligence-powered approaches for portfolio optimization. From the traditional methods of portfolio optimization, methods such as random matrix theory, shrinkage estimators, correlation asymmetries and partial correlation networks are presented. While, from the artificial intelligence realm, techniques such as machine learning efficient frontiers, performance-based regularization, neural network predictors and deep learning models for direct optimization of portfolio Sharpe ratio are highlighted. Intertwining the traditional methods, with artificial intelligence techniques, this review highlights relevant portfolio optimization research useful for academics and practitioners alike.

The expansion of the manufacturing processes network requires algorithms that can enable better planning and optimization of the manufacturing processes. Therefore, in the recent years the developments within the machine-learning (ML) and artificial intelligence (AI) have led to a new terminology, the so-called Industry 4.0. The fastest growth of Industry 4.0 has been encountered in the manufacturing, supply chain, services and products. The machine learning is prone to enable the development of smart supply-chains and manufacturing processes. The present research concerns the suitability and efficiency of the machine learning algorithm for the enhanced supply chain in manufacturing processes. The results show that the machine learning algorithm enables and enhances the efficiency of the manufacturing processes by clustering the machine-tools and increasing the number of manufactured components at the same tool location.

The article is devoted to the stylometric analysis of stories written by Jorge Luis Borges and Adolfo Bioy Casares under the common pseudonym Honorio Bustos Domecq. The work poses two questions: 1) Is Bustos Domecq’s style different from the writing style of Borges and Bioy Casares? 2) What is the share of influence of each of the co-authors on the formation of Bustos Domecq's style? To solve research problems, it is used Delta method, one of the most reliable stylometric tools to date; as well as the support vector machine, a common machine learning method that is used to solve classification problems. It turns out that Bustos Domecq's style differs from the style of the stories of Borges and Bioy Casares. However, in the texts of Bustos Domecq the authorial signal of Bioy Casares predominates, which is revealed on the basis of both stylometric and historical-literary analysis. The influence of Borges is more clearly manifested only in the first stories (1940s), when Bioy Caceres is a literary disciple of Borges, and in the second half of the 1960s, at a time of serious emotional upheavals for Borges, as well as his world recognition. Analysis of the topics of the stories, where Borges's authorial signal predominates, also confirms the results of stylometric experiments: the story “The Long Search for Tai An”, as well as the first texts of the collection The Chronicles of Bustos Domec “Tribute to Cesar Paladion”, “An Evening with Ramon Bonavena” and others really correspond to the poetics of Borges.

Accurate estimation of the elastic modulus (Mr) in the com- pacted subgrade soil is essential for the design of flexible pavement systems that are both reliable and environmentally friendly. Mr significantly affects the structural integrity of the pavement, especially in hilly areas with varying loads and climatic conditions. This study collects 2813 data points from pre- vious research results to create an accurate prediction model. The gradient boosted (GB) machine learning (ML) approach is selected to predict the Mr of the compacted subgrade soil. The accuracy and predictive performance of the GB model were evaluated using statistical analysis that includes fun- damental metrics such as root mean square error, mean absolute error, and relative squared error. The model obtained R² values of 0.96 and 0.94 for the training and testing datasets. The RMSE was 5 MPa for training and 7.48 MPa for testing, while the MAE was 3.18 MPa and 5.55 MPa. These results highlight the potential of GB in predicting soil Mr, thereby supporting the development of more accurate and efficient Mr prediction, ultimately reduc- ing time and cost.

Accurately forecasting the energy consumption of industrial equipment and linking these forecasts to equipment health has become essential in modern manufacturing. This capability is crucial for advancing predictive maintenance strategies to reduce energy consumption and greenhouse gas emissions. In this study, we propose a hybrid model that combines Long Short-Term Memory (LSTM) for energy consumption prediction with a statistical change-point detection algorithm to identify significant shifts in consumption patterns. These shifts are then correlated with the equipment’s health status, providing a comprehensive overview of energy usage and potential failure points. In our case study, we began by evaluating the prediction model to confirm the performance of LSTM, comparing it with several machine learning models commonly used in the literature, such as Random Forest, Support Vector Machines (SVM), and GRU. After assessing different loss functions, the LSTM model achieved the strongest prediction accuracy, with an RMSE of 0.07, an MAE of 0.0188, and an R2 of 92.7%. The second part of the model, which focuses on detecting change points in consumption patterns, was evaluated by testing several cost functions combined with binary segmentation and dynamic programming. Applied to a real-world case, it successfully detected a change point two months before equipment failure, offering the potential to reduce energy consumption by 27,052 kWh. This framework not only clarifies the relationship between equipment health and CO2 emissions but also provides actionable insights into emission reduction, contributing to both economic and environmental sustainability.

The growing urgency of environmental challenges and the depletion of fossil fuels have accelerated the search for sustainable and renewable energy sources. Wind energy, for example, is an important source of green electricity. However, using wind power is challenging due to the variability and unpredictability of wind patterns. Consequently, the ability to predict wind power in advance is crucial. The integration of artificial intelligence within the renewable energy sector could provide a viable solution to this challenge. In this study, we investigate the potential of machine learning to improve wind power forecasting by conducting a comparison of three regression models: K-Nearest Neighbor regression, Random Forest regression, and Support Vector regression. These models are combined with a feature selection technique to forecast wind power. Additionally, we propose a novel hybrid approach that combines these machine learning models with Multiple Linear Regression to address the complexities of wind energy forecasting. The performance of the models is evaluated using the R² score, Mean Absolute Error, and Root Mean Squared Error. The dataset for this study was generated from a numerical simulation conducted at a location with a latitude of 22.55° N and a longitude of -14.33° E. The findings demonstrate that the proposed hybrid model outperforms the individual machine learning models in terms of prediction accuracy. This study provides a solid foundation for future research and development in wind energy forecasting.

The present research focuses on solar radiation prediction, which is important for energy production in thermal and solar systems. For this purpose, open-source software (Python) and a methodology involving the creation, implementation, and testing of specific machine learning models random forest (RF) and decision tree (DT) were used. The metrics used to identify the effectiveness of the models in predicting solar radiation were the coefficient (R2), the mean square error (MSE), and the mean absolute error (MAE). The evaluation of the two methods is presented in three cases: for one, two, and seven days. The results show that the RF model has better results than the DT, with MAE and MSE values of 36.96 and 4238.77, respectively, and a determination coefficient of 0.96. The study emphasizes the importance of selecting the appropriate model based on the prediction horizon to estimate solar availability and improve solar and thermal energy system planning.

The study conducted focuses on predicting the different types of judicial cases presented to Moroccan administrative courts by using court decisions in the form of non-searchable PDF documents in the Arabic language. To achieve this, we utilized image processing, text cleaning techniques, and machine learning algorithms.We carried out a comparative study using both machine learning and deep learning techniques. The experiment was conducted in two phases: first on 697 court decisions, and then on 14,207 decisions from the Administrative Court of Appeal in Marrakech. Despite the challenges associated with the Arabic language, our methods were able to efficiently extract text, leading to accurate predictions. For the experiment on 697 decisions, machine learning achieved an accuracy rate of 91%, while deep learning reached 100%. For the experiment on 14,207 decisions, machine learning obtained an accuracy of 97%, and deep learning achieved 96%.As a result, this study contributes to the existing literature on the digitization and processing of unstructured documents in the Arabic language, as well as on the prediction of judicial case types through the use of machine learning and deep learning algorithms.

Passenger comfort is one of the critical factors in the railway system. Passenger comfort plays an important role in the success and effectiveness of railway transportation systems in terms of passenger satisfaction, health and well-being, economy, competition capability, and safety. For the railway infrastructure, one of the important components is a switch. Switches in the railway system have the function of guiding rolling stocks or trains to the preferred directions and tracks. However, switches are the components of the railway infrastructure that can negatively affect railway passenger comfort due to their geometry and the functions that are used to change the direction of rolling stocks. This study aims to study the relationship between the deterioration of railway switch support and railway passenger comfort by using machine learning. The data used in the study are axle box accelerations that are numerically simulated from verified multi-body simulation models. The machine learning technique used in the study is the convolutional neural network. The indicator used to evaluate the machine learning model’s performance is the accuracy. From the machine learning model development and training, the accuracy of the machine learning model is higher than 80% which is satisfied. Railway operators can benefit from the study’s findings by applying the developed machine learning model to collect data to evaluate the deterioration of railway switch support and the effect on railway passenger comfort.

Artificial intelligence (AI) is omnipresent in our daily lives and generates questions about its future place in medicine, in general, and in hepatology in particular. Numerous articles have been published in recent years on the subject, highlighting the potential role of AI, particularly in radiology for the diagnosis of hepatocellular carcinoma and hepatic dysmorphia, as well as in pathology. The use of AI will probably expand and play an increasingly important role at different stages of patient care: screening for cirrhosis in general medicine, predicting the risk of developing advanced fibrosis or esophageal varices. AI could also help identify the ideal candidate for a liver transplant for cirrhosis linked to an alcohol use disorder and interfere in the doctor-patient relationship by responding in a "more empathetic" way to questions asked by patients. This mini-review presents some applications of AI in our specialty. [ABSTRACT FROM AUTHOR]

The notion of artificial intelligence (AI) was born in the twentieth century, and many AI researchers have been promising us better days by integrating AI into our personal and professional lives. However, in recent years, as a result of technological developments, we have been faced with a massive intrusion of AI into our daily lives. AI allows us to analyse massive quantities of medical data in order to create algorithms to help diagnose and treat male infertility. However, the use of AI in andrology and more generally in the healthcare field also presents challenges. Confidentiality of medical data and legal liability in the event of algorithmic errors or biases are major concerns. In addition, it is essential to ensure that AI does not completely replace human medical expertise, but rather complements it to provide the best possible care for patients. [ABSTRACT FROM AUTHOR]

This article presents the work of urban corpus creation and processing, i.e. the methodology adopted, the choices made, the problems encountered and the results obtained at these two stages. It shows the work made in the context of the Vital project (Ville et traitement automatique des langues, “natural language processing and the city”), which draws from the digital humanities and brings together researchers in natural language processing (NLP), architects and urban planners in an original and innovative way. The aim of this multi-disciplinary research is to analyse the understanding of space as dynamic, using recent advances in the field of NLP.

This article outlines some of the challenges that have arisen during the first phases of the ERC-funded Modern project, a five-year research programme that takes a new “data driven” approach to the literary history of the French Enlightenment. Drawing on a large curated corpus of French texts of the Early Modern period, the authors describe in detail the various steps for building intertextual networks using the output of text reuse algorithms. From corpus and metadata cleaning to training a neural network for filtering ‘noisy’ passages, this article provides a pragmatic technical pipeline for similar projects working with massive collections of digitised text, highlighting both the promise and perils of conducting literary research at scale.

Nesta entrevista com a pesquisadora, artista e professora livre-docente Giselle Beiguelman (FAU-USP), são abordadas questões latentes disparadas pelas recentes tecnologias de Inteligência Artificial. Como pano de fundo das discussões, está o livro Políticas da Imagem: vigilância e resistência na dadosfera (2021), da própria entrevistada. São tratados temas como os processos artísticos e suas possibilidades frente às novas tecnologias, o debate ético (mas também estético) em torno dos deepfakes, a integração entre as visualidades e a política na contemporaneidade, a noção de preservação digital frente às mudanças tecnológicas e a viabilidade de se construir relações de aprendizado mútuo com as máquinas.

Objetivo: o objetivo deste artigo é apresentar uma revisão da literatura empírica recente a respeito do uso de métodos de screening (ou filtros econômicos) para a detecção de cartéis no Brasil e no mundo. Tais métodos têm sido cada vez mais utilizados por autoridades antitruste na medida em que complementam e reforçam métodos reativos tradicionais de detecção de cartel, como acordos de leniência. Há diversos estudos que discutem suas aplicações, a maioria com foco na abordagem de screening comportamental. Este artigo visa a complementar tais trabalhos de duas maneiras: (i) avaliar a literatura mais recente sobre o tema, a qual tem se especializado, por exemplo, em metodologias de machine learning; e (ii) discutir aplicações de abordagens estruturais, que são métodos mais simplificados de detecção do que o screening comportamental, mas também utilizados por autoridades antitruste e na literatura. Método: revisão e análise qualitativa da literatura sobre detecção de cartel nacional e internacional. Conclusões: diversas autoridades antitruste têm seguido as recomendações de organizações internacionais de combinar o uso de métodos reativos com proativos na detecção de cartéis. Há iniciativas importantes no uso de screenings, especialmente no âmbito de licitações públicas dada a maior disponibilidade de dados. Na literatura empírica técnicas de machine learning têm ganhado espaço seja devido aos desafios de identificação dos modelos econométricos, seja pelo contexto de aumento na complexidade do funcionamento dos acordos colusivos. Trata-se de uma nova tendência que está voltada principalmente à maior previsibilidade e precisão no processo de detecção de cartéis, em detrimento de conclusões sobre causalidade. Tendo em vista as diferentes dinâmicas setoriais e regionais na economia, sua utilização ainda exige cautela por parte dos pesquisadores e autoridades antitruste, mas certamente representa avanço relevante e promissor no combate a cartéis.

A temporal upscaling study was conducted to estimate net ecosystem exchange (NEE) of carbon dioxide and net methane exchange (NME) for a low-center polygon (LCP) ecosystem in the Mackenzie River Delta, for each of the 11 growing seasons (2009–2019). We used regression models to create a time series of flux drivers from in situ weather observations (2009–2019) combined with ERA5 reanalysis and satellite data. We then used neural networks that were trained and validated on a single growing season (2017) of eddy covariance data to model NEE and NME over each growing season. The study indicates growing season NEE was negative (net uptake) and NME was positive (net emission) in this LCP ecosystem. Cumulative carbon (C) uptake was estimated to be −46.7 g C m−2 (CI95% ± 45.3) per growing season, with methane emissions offsetting an average 5.6% of carbon dioxide uptake (in g C m−2) per growing season. High air temperatures (>15 °C) reduced daily CO2 uptake and cumulative NEE was positively correlated with mean air growing season temperatures. Cumulative NME was positively correlated with the length of the growing season. Our analysis suggests warmer climate conditions may reduce carbon uptake in this LCP ecosystem.

Linear model reduction techniques design offline low-dimensional subspaces that are tailored to the approximation of solutions to a parameterized partial differential equation, for the purpose of fast online numerical simulations. These methods, such as the Proper Orthogonal Decomposition (POD) or Reduced Basis (RB) methods, are very effective when the family of solutions has fast-decaying Karhunen–Loève eigenvalues or Kolmogorov widths, reflecting the approximability by finite-dimensional linear spaces. On the other hand, they become ineffective when these quantities have a slow decay, in particular for families of solutions to hyperbolic transport equations with parameter-dependent shock positions. The objective of this work is to explore the ability of nonlinear model reduction to circumvent this particular situation. To this end, we first describe particular notions of nonlinear widths that have a substantially faster decay for the aforementioned families. Then, we discuss a systematic approach for achieving better performance via a nonlinear reconstruction from the first coordinates of a linear reduced model approximation, thus allowing us to stay in the same “classical” framework of projection-based model reduction. We analyze the approach and report on its performance for a simple and yet instructive univariate test case.

Os problemas de saúde mental têm vindo a aumentar em todo o mundo, o que poderá estar associado ao crescimento da população urbana e ao estilo de vida a ela associado. O reconhecimento de que os diversos aspetos do ambiente ur-bano podem afetar a saúde mental dos indivíduos tem vindo a aumentar, uma vez que, estes são responsáveis por facilitar ou inibir comportamentos e estilos de vida que impactam o sentimento. Neste contexto é importante compreender o potencial impacte do ambiente urbano da cidade de Lisboa. Para tal recorreu-se à análise de sentimentos, utilizado um léxico do NRC Sentiment and Emotion, a partir de dados da rede social Twitter, possibilitando a identificação dos locais em que tanto o sentimento positivo como negativo prevalecem. De seguida fez-se uso de um modelo de machine learning (ML) associado a um modelo-agnóstico de modo a aumentar a compreensão dos fatores do ambiente urbano que podem explicar o sentimento. Foram testados 4 modelos de ML, Random Forest (RF), Ex-treme Gradient Boosting (XGBoost), Neural Network (NN), o K-Nearest Neighbour (KNN) e um modelo linear para comparação (Generalized Linear Model - GLM). Os modelos agnósticos aplicados, o Local Interpretable Model-Agnostic Explanations (LIME) e o SHapley Additive exPlanation (SHAP), desempenharam um papel funda-mental neste estudo. Respondendo à questão de partida, as variáveis explicativas que mais se relacionam com o sentimento são a distância a equipamentos fitness, a distância aos espaços verdes, a popularidade dos locais (estimada através da rede social Flickr) e a distância à rede ciclável.

In this paper, we applied the artificial intelligence technique (SVM Classifier) to compare the performance of two different technologies of PV modules (class to class and backsheet to glass) after five (05) months of operation in Algeria under the same weather conditions (moderate and humid climate) . We have a database for the outdoor monitoring of these two PV modules, consisting of data (Isc, Voc, Pmax, Imp, Vmp, Tm, Tamb, G, WD, WS, Date, Time) which are variables data, where the SVM creates the groups or class according to the conditions that we entered, after which it produces heatmaps that help us in reading the results and making the decision easily, unlike the classic methods which are very difficult. This method is applicable for comparison between several solar panels or several photovoltaic PV plants. It is enough just to give the database.

Many complex chemical problems encoded in terms of physics-based models become computationallyintractable for traditional numerical approaches due to their unfavorable scaling with increasing molecular size. Tensor decomposition techniques can overcome such challenges by decomposing unattainably large numerical representations of chemical problems into smaller, tractable ones. In the first two decades of this century, algorithms based on such tensor factorizations have become state-of-the-art methods in various branches of computational chemistry, ranging from molecular quantum dynamics to electronic structure theory and machine learning. Here, we consider the role that tensor decomposition schemes have played in expanding the scope of computational chemistry. We relate some of the most prominent methods to their common underlying tensor network formalisms, providing a unified perspective on leading tensor-based approaches in chemistry and materials science.

Excelzyme, an enzyme engineering platform located at the Zurich University of Applied Sciences, is dedicated to accelerating the development of tailored biocatalysts for large-scale industrial applications. Leveraging automation and advanced computational techniques, including machine learning, efficient biocatalysts can be generated in short timeframes. Toward this goal, Excelzyme systematically selects suitable protein scaffolds as the foundation for constructing complex enzyme libraries, thereby enhancing sequence and structural biocatalyst diversity. Here, we describe applied workflows and technologies as well as an industrial case study that exemplifies the successful application of the workflow.

Artificial Intelligence (AI) represents the ability that technologies or machines have to copy human intelligence as close as possible in order to solve problems and achieve goals. Artificial intelligence systems adapt, analyze data, observe future actions based on existing information and operate autonomously. An interesting change has occurred over time. In the past, the focus was on the hardware, while the software was considered a weak element. Over time, the software element developed, and over time hardware engineers adapted to the evolution becoming software engineers. Algorithms are used to make predictions in almost any field, and if used correctly, the predictions and results are beneficial and commendable. The take-up of these artificial intelligence applications in public institutions is useful to all. Therefore, developing and perfecting basic human skills is important in the long run. Above all, technology enables work to become more human. For managers, leaders or directors it has a tremendous result. It should be pointed out that starting from the first light bulb up to the emergence of the smartphone, technology has evolved. The element that never changes is the people behind the technology, while the most important aspect is that artificial intelligence is changing the working world.

This paper revisits the different arts of engineering. The art of modeling for describing the behavior of complex systems from the solution of partial differential equations that are expected to govern their responses. Then, the art of simulation concerns the ability of solving these complex mathematical objects expected to describe the physical reality as accurately as possible (accuracy with respect to the exact solution of the models) and as fast as possible. Finally, the art of decision making needs to ensure accurate and fast predictions for efficient diagnosis and prognosis. For that purpose physics-informed digital twins (also known as Hybrid Twins) will be employed, allying real-time physics (where complex models are solved by using advanced model order reduction techniques) and physics-informed data-driven models for filling the gap between the reality and the physics-based model predictions. The use of physics-aware data-driven models in tandem with physics-based reduced order models allows us to predict very fast without compromising accuracy. This is compulsory for diagnosis and prognosis purposes.

The rapid evolution of technologies in the field of justice has led to the creation of artificial prediction tools that can be used to predict the outcome of disputes and support judges in their decisions. However, the use of these predictive justice tools has been debated due to concerns regarding the validity and reliability of prediction models and their impact on the principle of fairness in the justice system. In this article the impact of predictive justice systems on the decisionmaking of judge will be examined.

This paper presents a comprehensive study of a predictive irrigation system, an innovative approach in smart agriculture focusing on integrated irrigation management through advanced predictive techniques. Employing a blend of Internet of things (IoT) technology, Machine Learning (ML) algorithms, and data analytics, this system marks significant improvements in agricultural irrigation strategies. It is designed to optimize water use, improve crop yields, and promote sustainable farming practices in the face of evolving environmental challenges. The paper outlines the system's architecture, including the deployment of IoT sensors for continuous data collection, the integration of ML models for predictive analysis, and the implementation of adaptive irrigation scheduling algorithms. A detailed examination in a study case of the system's performance reveals substantial improvements in water usage efficiency compared to traditional irrigation methods. Additionally, the paper discusses the challenges and limitations encountered, such as the high initial setup costs, technical complexities, and the necessity for continuous data accuracy. The study concludes by underscoring the Irrigation predictive system's potential in transforming agricultural practices. It highlights its role in enhancing resource management and sustainability in farming, while also pointing out the areas for future research to further refine of system for wider applicability.

Prediction is widely used for various purposes and in many fields of human activity. The techniques employed for making predictions are a subject of great scientific interest within the research community due to their diversity, level of accuracy, and adaptability to data. The challenge is to determine the factors that affect the choice of an optimal technique suited to each prediction objective. In this article, we conduct a review of models used in the literature to make predictions in different domains to understand the factors influencing the selection of a specific predictive model in relation to their areas of study. A comparative analysis of prediction techniques such as statistical algorithms, Data Mining, and Machine Learning has been performed. It follows that the selection of an adequate prediction technique for the best decision-making should take into account the projection horizon, uncertainty around the prediction, data availability and reliability, and the associated cost of prediction.

This paper presents IEML (Information Economy MetaLanguage), a uniform representation of human meaning and knowledge that can be read and processed automatically by machines. Distinguished from pragmatic and referential semantics, linguistic semantics is today incompletely formalised. Only its syntagmatic dimension has been mathematised in the form of regular languages. Its paradigmatic dimension remained to be formalised. To solve the problem of the complete mathematising of language, including its paradigmatic dimension, I propose to code the linguistic meaning in IEML. IEML has the same expressive capacity as a natural language and has an algebraic structure allowing the calculation of its semantics. The article explains its dictionary, its formal grammar, and its integrated tools for building semantic graphs. As far as its applications are concerned, IEML could be the vector of a fluid calculation and communication of meaning – semantic interoperability – capable of decompartmentalising the digital memory and feeding the progress of collective intelligence, artificial intelligence, and digital humanities. I conclude by indicating some research directions. This paper presents a synthesis of several decades of research.

Artificial intelligence (AI) applied to accounting and auditing represents a matter that is getting the attention of researchers. This paper provides an understanding of the evolution of AI and its role in shaping accounting information systems (AISs). It aims to summarize the current state of the literature that deals with the impact of AI on AIS, highlighting related aspects, such as the benefits and challenges of AI in AIS research. This review suggests that current studies dealing with this matter are scarce and invites future research to examine how bias and transparency are handled in the context of AI auditing systems and whether human auditors will fully rely on AI outcomes.

Copyright of BMS: Bulletin de Methodologie Sociologique (Sage Publications Ltd.) is the property of Sage Publications Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In this study, structure – property relationships (SPR) have been investigated using machine learning methods (ML). The research objective was to create a ML model that can predict the stress – strain response of materials at different temperatures from a given microstructure with industrially acceptable accuracy and high computational efficiency. Automated microstructure generation techniques and numerical simulations were developed to create a dataset for the ML model. Two – phase 3D representative volume elements (RVEs) were analyzed using finite element analysis (FEA) to obtain the stress – strain responses of the RVEs. The phase arrangement of the RVEs, the temperature, and the stress – strain responses were used to train the ML model. The microstructure arrangement and the temperature – dependent mechanical properties of each phase were known parameters, while the output parameter was the stress – strain response of the two – phase RVE. The ML model has shown excellent prediction accuracy in the temperature range from 20 °C to 250 °C. In addition, the model showed very high computational efficiency compared to FEA, allowing much faster prediction of the stress – strain curves at specific temperatures.

Today, artificial intelligence has become a common concept in many fields of technical sciences and humanities, the presence of smart phones in our hands is the best proof of that, as it became necessary to acquire smart devices and deal with smart information programs, Artificial Intelligence is also considered one of the most successful fields at the present time, as it came out of the research phase into the field of development and employment.Artificial intelligence began to take center stage when it was considered a scientific breakthrough during the past two decades due to the superior skills and achievements that resulted out of it in various fields, including medicine, commercial, industrial and educational business organizations, and security systems, Artificial intelligence or machine learning relies on algorithms to simulate the human learning process, as it enters in all areas that need logical thinking, knowledge, planning and hypothetical awareness based on applying theories and choosing the right solutions, and marketing is one of these area where AI is driving some of the current biggest technological breakthroughs in marketing and advertising.The field of marketing and advertising has been affected by artificial intelligence, so adopting the growth and development of artificial intelligence is crucial for marketing and advertising efforts, where institutions use artificial intelligence programs to improve their efficiency, reduce costs, and improve their mental image among the recipients, which helped in developing the designs and operational structures of the institutions.The key to successful marketing campaigns is through a qualitative analysis of recipient data and the study of the target group, although artificial intelligence is used by a wide range of applications in many scientific fields, the amount of data that is collected daily provides the opportunity to analyze, design and implement marketing strategies aimed at developing decision-making patterns based on smart knowledge, The effectiveness of artificial intelligence applications is represented in reducing the potential risks of advertising campaigns and competitors by analyzing many data and carrying out statistics to reduce error rates.The research is concerned with emphasizing the vital role that artificial intelligence plays in making marketing more intelligent, impactful and relevant to the recipient, Therefore, during the coming period, artificial intelligence is expected to turn into a basic and effective feature in order to enhance the presence and competition of advertising

Approaching to PDCA (Plan - Do - Check - Take Action) in management of infrastructure asset requires digital transformation, sufficient data and strong database supporting management, analysis as well as creation of data-driven decision making tools. For pavements, data including condition indicators such as roughness and rutting depth are collected automatically during the survey vehicle travelling. However, pavement crack ratio and crack features of pattern and segmentation have not been detected by the system but manual in the case in Vietnam. The paper presents result of research on algorithm of statistic machine learning model in AI applying deep learning algorithm to automatically detect crack feature on pavement photos for enhancement of the performance and productivity of current survey technology. In the research, a deep architecture using convolutional neural network (CNN) for crack segmentation on gray scale images has been developed. The results show the CNN model for crack segmentation is better than other methods using traditional digital processing such as the Gabor filters or threshold and machine learning such as Adaboost.

Machine learning and essentialism have been connected in the past by various researchers, in order to state that the main paradigm in machine learning processes is equivalent to choosing the “essential” attributes for the machine to search for. Our goal in this paper is to show that there are connections between machine learning and essentialism, but only for some kinds of machine learning, and often not including deep learning methods. Similarity-based approaches, more connected to the overall prototype theory, spanning from psychology and linguistics, seem more suited for pattern recognition and complex deep-learning issues, while for classification problems, mostly for unsupervised learning, essentialism seems like the best choice. In order to illustrate the difference better, we will connect both paths to their sources in other disciplines and see how human psychology influences our decision in machine-learning modeling as well. This leads to a philosophically very interesting consequence: even in the setting of supervised machine learning, essences are not present in data, but in targets, which in turn means that the categories which purport to be essences are in fact human-made, and hand-coded in the targets. The success of machine learning, therefore, does not give any substantial evidence for the independent existence of essential properties. Our stance here is to state that neither the existence nor the lack of “essential” properties in machine learning can lead to metaphysical, i.e., ontological claims.

This paper explores a series of thoughts about the meaning of intelligence in neuroscience and computer science. This work aims to present an understandable definition that fits our contemporary artificial intelligence background. The research methodology of this essay lies in existing theories of artificial intelligence, focused on computer science and neuroscience. I analyze the relationship between intelligence and neuroscience and Hawkin’s Thousand Brains Theory, an approach to show what it is an intelligent agent according to neuroscience. Here, the main result relies on the verification that intelligence is only possible in the neocortex. According to this result, the study performs a second critical analysis aiming to demonstrate why there is no artificial intelligence today.

Copyright of Vie et Sciences de l'Entreprise is the property of ANDESE and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

A presente pesquisa parte do atual contexto da revolução industrial 4.0 e popularização do uso de inteligências artificiais em diversos eixos da vida em sociedade, mais especificamente no desenvolvimento de carros autônomos. O problema da pesquisa visa responder quais as possíveis repercussões da responsabilidade civil envolvendo inteligências artificiais envolvendo carros autônomos podem ocorrer no Direito do Consumidor. O objetivo geral visa explorar as perspectivas sobre as Inteligências Artificiais e Veículos autônomos, bem como as possíveis repercussões dentro da responsabilidade consumerista. Especificamente, objetiva-se esclarecer sobre o estado da arte das inteligências artificiais; explanar sobre as expectativas, níveis de autossuficiência, acidentes e possíveis falhas dos carros autônomos; descrever a responsabilidade no caso de dano no Direito do consumidor; e esboçar pontos em que as novas tecnologias provocam dificuldades e talvez um repensar da responsabilidade civil no direito do consumidor. A metodologia usada parte de pesquisa predominantemente teórica, embora utilizando-se elementos de empiria em análise de casos; com métodos de natureza básica e objetivos exploratórios e abordagem qualitativa. Quanto ao procedimento, o método utilizado foi o de levantamento bibliográfico-documental, aplicando-se majoritariamente a lógica indutiva. Conclui-se que os fornecedores devem desenvolver estratégias preventivas e proativas para evitar acidentes de consumo e debater sobre os princípios éticos a serem adotados pelas máquinas, bem como que as inteligências artificiais têm um grande impacto na seara jurídica, especificamente no direito do consumidor, sendo necessário reinterpretar conceitos como "defeito", “produto” e “serviço”, que repercutem nas excludentes de responsabilidade e nas teorias do nexo causal.

The RXN for Chemistry project, initiated by IBM Research Europe – Zurich in 2017, aimed to develop a series of digital assets using machine learning techniques to promote the use of data-driven methodologies in synthetic organic chemistry. This research adopts an innovative concept by treating chemical reaction data as language records, treating the prediction of a synthetic organic chemistry reaction as a translation task between precursor and product languages. Over the years, the IBM Research team has successfully developed language models for various applications including forward reaction prediction, retrosynthesis, reaction classification, atom-mapping, procedure extraction from text, inference of experimental protocols and its use in programming commercial automation hardware to implement an autonomous chemical laboratory. Furthermore, the project has recently incorporated biochemical data in training models for greener and more sustainable chemical reactions. The remarkable ease of constructing prediction models and continually enhancing them through data augmentation with minimal human intervention has led to the widespread adoption of language model technologies, facilitating the digitalization of chemistry in diverse industrial sectors such as pharmaceuticals and chemical manufacturing. This manuscript provides a concise overview of the scientific components that contributed to the prestigious Sandmeyer Award in 2022