• 2015-2024close
• Open Access close
• Embargo close
• English close
• Digital Humanities and Cultural Her... close
• Knowmad Institut close

Ever since the explosion of the internet, fake news has always been a cause for concern. The proliferation of fake news online hinders access to reliable information. The efficiency of several machine learning methods for the identification of fake news is investigated in this work. We train and evaluate five models: Support Vector Machine (SVM), Logistic Regression, Random Forest, Long Short-Term Memory (LSTM), and Naive Bayes. Employing two distinct datasets, we evaluate the models' generalizability. We extract textual features from the news articles and assess their performance using established metrics. This investigation sheds light on the advantages and limitations of each model within the context of fake news classification, contributing to the development of more robust detection systems. Furthermore, we explore the impact of utilizing different machine learning paradigms, including supervised learning (Logistic Regression, Random Forest, SVM) and deep learning (LSTM) on the detection accuracy. This comparative analysis provides valuable insights into the optimal approach for tackling the intricate challenge of fake news identification.

ES CIENCIA is a research platform created by the Spanish National Research Council (CSIC) in 2019 to work on Spanish as a language of scientific communication. It develops interdisciplinary research projects in inter-institutional and international collaboration, both with Latin America and Europe. ES CIENCIA involves researchers and technical specialists in Academic Publishing, Terminology, Information& Library Science, Computational Linguistics, Artificial Intelligence and Natural Language Processing and Translation.The apparent oxymoron in the title is intended to draw attention to the fact that work on a particular language in science, and in the context of open science, contributes to the development of a multilingual scientific communication system. In fact, the platform is based on the values of multilingualism in science. ES CIENCIA has promoted the adhesion of the CSIC to the Helsinki Initiative on multilingualism in science communication, the participation of Spain in OPERAS, and has participated in the document on scientific publications read to the Ibero-American Ministers of Science, Technology and Innovation (2022) to highlight the value of the languages of the region in the transfer of research results to society.This poster will show the different levels at which ES CIENCIA works, the disciplines involved, the transfer of results and the interaction with institutions working in science, language and book policies. The main research projects will also be presented: TeresIA. Artificial and natural intelligence-based project on terminology in Spanish, Cartography of the IberoAmerican academic book publishing, Scholarly Publishers Indicators. Information and indicators about book publishing, Academic book publishing in Spain: Digital transformation, Open access books in Spain (within the OPERAS project PALOMERA.Policy Alignment of Open Access Monographs in the European Research Area), Name Entity Recognition in academic books in Spanish.

The Abstractive News Captions with High-level cOntext Representation (ANCHOR) dataset contains 70K+ samples sourced from 5 different news media organizations. This dataset can be utilized for Vision & Language tasks such as Text-to-Image Generation, Image Caption Generation, etc.

This paper discusses the critical disconnect between displays of religiosity and morality in contemporary churches in Nigeria. The paper examines whether religiosity has enhanced the spiritual and moral formation of Christian adherents in terms of curbing corruption as a major moral cancer that affects all aspects of human life. The writer subsumes all the moral issues under corruption. Some Nigerian churches are bedevilled by corrupt practices such as greed, fraud, adultery, fornication, and other moral vices against which they preach. Pastors embezzle church funds, engage in financial fraud, and engage in all forms of sexual immorality and abuse. The author used books, journals, bulletins, and internet articles relevant to the study. The findings of the study revealed five major indicators or barometers of spirituality through which adherents of the Christian religion in Nigeria display religiosity: pilgrimage, worship, regular attendance, fasting, and media production. This paper argues that these displays of religiosity have not been translated into the spiritual and moral formations of Christian religious adherents. The paper concludes by recommending to all Christians and churches in Nigeria the need to return to the practice of ethical and moral codes contained in the Bible, the theopneustic book of Christians.

[ES] Últimamente he estado trabajando como becario en MAHLE Electronics SLU. Allí descubrí lo ineficiente y tediosa que es la gestión de la información técnica, y no se limita a esta única empresa. Para resolver este problema, deberíamos encontrar una forma más eficiente de almacenar información o pasar horas memorizando archivos PDF para no tener la necesidad de buscarlos y, cuando se cambie una determinada norma, hacer todo el proceso otra vez. Hay demasiadas regulaciones, con demasiadas variables, se actualizan constantemente cada año y no existe un sistema para organizar toda esta información más allá del anticuado explorador de archivos. Por eso vamos a proponer una solución alternativa. Esta solución se basa en la idea de un asistente de inteligencia artificial personalizado que extrae información de una base de datos predeterminada. La razón por la que esto es necesario es que los motores de búsqueda habituales están demasiado repletos de información y no son muy específicos cuando se trata de información técnica. Además, no pueden acceder a archivos privados o confidenciales que contengan los datos necesarios. El objetivo es crear un asistente de inteligencia artificial que utilice algoritmos de knowledge retrieval para extraer y sintetizar datos de una colección de archivos PDF. La forma de interactuar con este asistente será a través de un chatbot de Telegram. Por lo tanto, necesitaremos crear el asistente con la información y luego vincularlo a un chatbot de Telegram para que podamos comunicarnos con él. Para construir el asistente usaremos el código Python proporcionado por OpenAI y su API asistente, luego lo vincularemos al chatbot de Telegram a través del ID del bot. Todo esto se hará en código Python. Finalmente, evaluaremos el rendimiento de este bot en la búsqueda y síntesis de información respecto a un trabajador de MAHLE Electronics SLU. Este TFG evalúa el desempeño del bot, su viabilidad y su potencial. Los resultados muestran cómo un espacio de trabajo podría beneficiarse con la inclusión de este bot. [EN] Lately, I have been working as an intern in MAHLE Electronics SLU. There I discovered how inefficient and tiresome the management of technical information is, and it s not limited to this one company. To solve this issue, we should come up with a more efficient way of storing information ore spend hours memorizing PDFs so that we don t have the need to search for them and, when a certain regulation is changed, do it all again. There are too many regulations, with too many variables, they are constantly getting updated every year and there is not a system in place to organize all this information beyond the antiquated file explorer. That is why we are going to propose an alternative solution. To do so, we came with the idea of a personalized AI assistant that draws information from a pre-determined database. The reason this is needed is that regular search engines are too bloated with information and are not very specific when it comes to technical information. Also, they cannot access private or confidential files that contain the needed data. The objective is to build an AI assistant that uses knowledge retrieval algorithms to extract and synthesize data from a collection of PDFs files. The way to interact with this assistant will be through a Telegram chatbot. So, we will need to build the assistant with the information, then link the assistant to a telegram chatbot so that we can communicate with it. To build the assistant we will use the python code provided by OpenAI and its assistant API, then we will link it to the telegram chatbot through the bot s ID. All of this will be done in python code. Finally, we will evaluate the performance of this bot at finding and synthesizing information compared to a worker in MAHLE Electronics SLU. This bachelor s degree thesis evaluates the performance of the bot, its feasibility, and its potential. The results show how a general workspace could benefit by the inclusion of this bot.

This dissertation explores representations of Sweden and Denmark and how immigrants’ place in Sweden and Denmark are represented in instructional materials for adult learners in the period ca. 19602005. Denmark and Sweden are often pitted as opposites in terms of immigrant incorporation, Denmark representing a more assimilatory stance and Sweden that of a sanctioned multiculturalism. Instead of looking at policy, this dissertation looks at the incorporative messages meeting immigrants in an attempt to better understand the development of the topic of integration. Theoretically, the study relies on discussions of concepts of immigrant incorporation i.e. assimilation, integration, and segregation. However, this is supplemented heavily with theory on nationalism as an ideology. Specifically, A.D. Smith serves as inspiration and the analytical distinction between ethnic and civic identity is discussed throughout. Additionally, Billig´s concept of Banal Nationalism is discussed.The dissertation features an analysis of a time before any particularly regulated attempts at teaching immigrants were undertaken, and the materials produced at the time focused primarily on language acquisition. As state-organized language instruction for immigrants emerges, as well as a conceptualization of immigrants as permanent members of society, this analysis shows that the materials nationalize in the sense that they start to feature a crash-course of the compulsory education system. This crash-course in belonging remains as central to the instructional materials in the remainder of the time period. In an early phase of instruction, ca. 1965 to the early 1980s, representation of employment and genderequality are examined and the two cases are found to differ. The Swedish materials reflect a strong politicization of the immigrant reader, while no politicization is found in the Danish case. An examination of the representation of culture in the materials is done, finding no real difference between the Danish and Swedish case. To the 1970s, the sources feature a discussion of cultural difference was grounded in an idea of benign difference. They feature the idea that cultural contact itself was the problem. From the early 1980s, the materials reflect a new understanding of immigrant culture, and the idea that Sweden and Denmark are culturally affected by immigration. Representations of a need to change immigrants to a Swedish or Danish gender-equal norm, and a textually projected immigrant other at odds with labor market participation become common. At the same time, a shift in history used for the purpose of integration becomes common. Taken as a whole, the materials in this later phase are discussed as bearers of an inclusionary and exclusionary sentiment, reproducing both ethnic and civic tenets of national identity. This complexity is discussed as an important result of this dissertation.

The digital era's advancements have prompted the adoption of communication as the primary medium for the corporate industry. Formerly, business discussions, profiles, conferences, purchasing, and settlements were all carried out in person. But the modern era has made everything digitized. In the past few years, it's been observed there is an exponential increase in the count of complicated manuscripts and writings that need a better recognizing of machine learning methodologies to successfully detect languages in various purposes. Several Artificial Intelligent approaches have produced outstanding achievements in processing natural languages. The ability of various machine learning and deep learning to realize complex models and non-linear associations within data is critical to their efficiency. Learning applicable frameworks, architecture, and algorithms for input classification, such as text files, audio, and video files, on the other hand, is a challenging task. Objective: This study aims at Natural Language Processing in the identification of text, voice messages, smart records, and chatbots. Hybrid deep learning approach for the classification of the inputs that are in the form of text, voice, and video records. Problem Statement: As interaction becomes more crucial to business, firms have designed sophisticated NLP programs. These NLP take human wishes and satisfy them quickly through messages, telephone calls, digital records, and chatbots. The ease of communication and connection has shown a stronger impact on customer preferences, aspirations, and demands. Contemporary service providers today utilize email, messaging, telephone calls, digital records, and chatbots as primary points of contact for practically all of their transactions, client inquiries, and preferable trade channels. Method: The study uses text content, voice message, and audio as part of Natural Language Processing and Hybrid Deep Learning approaches to demonstrate how input is processed depending on user reactions, replies to text messages, and audio record identification during communications.

Risk management, being the pivotal function along various sectors, ranging from finance to healthcare, is going through profound changes with the idea generation phase of data science. The paper shows the juncture where risk management and data science meet, and the dramatic effect of data-driven solutions on decision-making and organizational adaptability are demonstrated. By utilization of approaches including predictive modeling, machine learning, and artificial intelligence, risk assessment, identification, and mitigation are more effective given the power of data volumes a business may have at its disposal. Data science presents a way of accomplishing an integrated and holistic investigation of risks through the integration of a diversity of data sources, like the structured and unstructured data, which leads to higher accuracy of risk assessment. Predictive analytics assist organizations in analyzing risks and their underlying drivers before they occur and coming up with preventative strategies. Real-time tracking not only improves risk management activities but also facilitates the quick detection of errors or deviations from standard trends given faster responses to emerging risks. Also NLP and sentiment analysis help companies to detect public opinion and preempt potential reputational threats, offering a proactive approach to reputation management. The possibilities brought by data science in risk management are remarkable and need to be dealt with despite the hurdling problems of data quality, privacy and model interpretability. It is crucial to use accurate and trusted input data for the risk assessment as well as addressing the problems related to visibility and responsibility in environment where decisions are made by algorithms. As technology in data science continues to improve, new opportunities arise for even greater effectiveness in risk management. Through the identification and mitigation of current problems as well as the recognition and adoption of new trends, companies have the opportunity to utilize data science in the face of the uncertainties of the world, aiding in the creation and prioritization of strategic decisions resulting in greater resilience and sustainability.

The integration of Artificial Intelligence (AI) into Building Information Modeling (BIM) has the potential to revolutionize the Architecture, Engineering, and Construction (AEC) industry by enhancing efficiency, improving decision-making processes, and making BIM models more accessible and intuitive for professionals across various disciplines. This thesis introduces DAVE (Digital Assistant for Virtual Engineering), a prototype of a Generative Pre-trained Transformer (GPT)-powered digital assistant designed to facilitate seamless, real-time interactions and updates within BIM environments through natural language processing (NLP) and voice commands. By integrating with Autodesk Revit using Python scripts, the Revit API, and the OpenAI API, DAVE demonstrates a practical application of conversational AI in the AEC sector, aiming to make BIM workflows more intuitive and reducing the cognitive load on users. The development and implementation of DAVE involve a comprehensive system architecture that combines Python scripting, a Dynamic-Link Library (DLL), and a JavaScript Object Notation (JSON) file for efficient data management and interaction with BIM models. The prototype's testing and validation phase highlights its capability to handle a variety of commands, from simple tasks such as undoing actions to more complex operations like updating room names and numbers. Through a detailed analysis of system performance and user interaction, this thesis explores the challenges and limitations of integrating AI with BIM, including scalability, data integrity, and the quality of user queries. Despite these challenges, DAVE's successful demonstration of real-time BIM model management through voice and natural language commands marks a significant step forward in the digital transformation of the AEC industry. This thesis not only contributes to the ongoing development of AI applications within BIM but also lays the groundwork for future research aimed at expanding the capabilities of AI-powered assistants in the AEC sector. By addressing the current system limitations and exploring avenues for further enhancements, such as interoperability with other BIM software and improving the efficiency of information retrieval, the study paves the way for a more connected, intelligent, and user-centric future in the built environment. The development of DAVE embodies the potential of conversational AI to revolutionize architectural and construction methodologies, offering insights into how the AEC industry can leverage AI to achieve greater efficiency, sustainability, and innovation in project management and execution.

Freed from Athenian tutelage in 314 BC, at a time of geopolitical changes that marked the beginnings of the Hellenistic period in the Aegean world, Delos gradually consolidated its political and economic independence. During the third and second centuries, the Delian community redefined the central place that the island had continually occupied in the economic, financial and cultural flows of the Mediterranean. This study, mainly based on epigraphic accounting sources, including more than five hundred accounts and engraved inventories that were displayed in the sanctuary of Apollo, but also on numismatic sources and archaeological remains on the seafront, re-considers the question of Delos’ place in the Hellenistic economy. Far from being an exception to be excluded from serialized comparisons, the Delian evidence is indicative of Aegean economic circumstances and demonstrates the capacities of the Greek communities to adapt to change in troubled times. Behind the numbers cut in stone appear human communities and societies whose economic activities shed fresh light on the history of this part of the Mediterranean.