Search

The emergence and development of conspiracy phenomena within the digital space is broadly studied in the literature. The spread of junk or misleading information online is amplified by the increasing use of Social Networks and social media. Specifically on Facebook, there are networks that are specialized in disseminating information from unofficial sources. The aim of this work is to create an agent-based computational model capable of describing and analysing the structure and the development of a conspiracy echo-chamber within Facebook. We map and analyse the development and characteristics of a conspiracy echo-chamber within Facebook, through the study of interactions of articles from unofficial sources within the social network. Then, we explore the variation of the network structure in relation to the variation of certain properties of the articles. A set of preferential sources of information emerge, as if they are somehow being credited by the conspiracy public as references for the dissemination of "informative" content. We also find that there is a polarized bubble of content and users with respect to conspiracy-mindedness. We show that there are pages that have a preferential role within the network as opinion leaders in the conspiracy bubble. Moreover, we also show that there are elements that distinguish many of the analyzed contents, namely political orientation and hate speech. Finally, we highlight that the grounded approach is very useful in the classification process because approaching classification without a-priori categories to be identified in the text make it possible for issues and specific interpretative codes to emerge.

Secondo il World Economic Forum, entro il 2025 almeno il 50% dei lavoratori in ogni settore necessiterà di significativi processi di re-skilling e up-skilling, mirati ad aumentare o aggiornare le competenze della forza lavoro. In questo scenario complessivo, docenti, formatori ed esperti, in qualsiasi ambito educativo, necessitano di sistemi di raccomandazione, basati su intelligenza artificiale, in grado di supportarli nella creazione di corsi e percorsi di apprendimento personalizzati. I docenti hanno bisogno di tool che li aiutino a ottenere i giusti suggerimenti sulle risorse didattiche e il materiale formativo da selezionare e integrare all' interno di corsi specifici che devono creare in base alle loro esigenze. Il mio progetto di dottorato executive è stato quindi finalizzato alla progettazione, realizzazione e valutazione di un sistema di raccomandazione explainable per la didattica a distanza, in grado di fornire ai docenti o formatori dei suggerimenti intelligenti e adeguatamente motivati per accompagnarli nella creazione di corsi online personalizzati. Il nostro obiettivo specifico è stato quello di realizzare un sistema pronto per il mercato per supportare docenti o formatori, potenzialmente in qualsiasi comunità e ambiente didattico. La soluzione proposta è stata pensata per migliorare e potenziare la piattaforma di e-learning "WhoTeach", prodotta e commercializzata da Social Things srl e concepita come un Social Intelligent Learning Management System. Il progetto è stato motivato principalmente da esigenze aziendali e di mercato, oltre che da una prima valutazione della piattaforma, ovvero un'attività di sperimentazione per valutare il livello di usabilità e la qualità dell'interazione tra gli utenti e un "prototipo dimostrativo" di sistema di raccomandazione di WhoTeach, da cui è emersa la forte necessità di sviluppare un sistema di raccomandazione explainable e potenziato. Successivamente, tenendo conto anche delle esigenze dell'azienda di, According to the World Economic Forum, by 2025 at least 50% of all employees will require significant re-skilling and up-skilling processes, aimed at increasing or updating the skills of the workforce. This need is definitively perceived by the companies stakeholders: according to Boston Consulting Group in 2019 , 67% of employees are likely to learn new skills in order to qualify for new positions or jobs, under any circumstances. Over the two last pandemic years, this trend grew stronger: in 2021 Boston Consulting Group reports that more than two-thirds of workers globally are willing to retrain for new jobs. In this overall scenario, teachers, trainers and experts, in any educational domain, need artificial intelligence based recommender systems able to support them in the creation of personalized courses and learning paths. Teachers require AI tools to get the right suggestions on the didactical resources to be selected and integrated into specific courses they have to create according to their requirements. Explainable approaches have the capability to allow teachers to understand the reason behind specific recommendations, thus improving their trust and collaboration with the system and, eventually, the creation of the courses. Then, my PhD project was aimed at designing, developing and evaluating an Explainable Recommender System solution for Educational Platforms, able to provide teachers with intelligent and explainable suggestions to support them in the creation of personalized online courses, in potentially any didactical community and environment. The proposed solution was aimed to improve and empower the educational platform “WhoTeach”, produced and commercialized by Social Things srl and conceived as a Social Intelligent Learning Management System. The project was primarily motivated by both company and market needs, as well as by an initial assessment of the platform, namely an experimentation activity to assess the level of usability and quality of i

Objective: To assess the efficacy of corticosteroids in patients with coronavirus disease 2019 (COVID-19). Methods: A multicentre observational study was performed from 22 February through 30 June 2020. We included consecutive adult patients with severe COVID-19, defined as respiratory rate ≥30 breath per minute, oxygen saturation ≤93% on ambient air or arterial partial pressure of oxygen to fraction of inspired oxygen ≤300 mm Hg. We excluded patients being treated with other immunomodulant drugs, receiving low-dose corticosteroids and receiving corticosteroids 72 hours after admission. The primary endpoint was 30-day mortality from hospital admission. The main exposure variable was corticosteroid therapy at a dose of ≥0.5 mg/kg of prednisone equivalents. It was introduced as binomial covariate in a logistic regression model for the primary endpoint and inverse probability of treatment weighting using the propensity score. Results: Of 1717 patients with COVID-19 evaluated, 513 were included in the study, and of these, 170 (33%) were treated with corticosteroids. During hospitalization, 166 patients (34%) met the criteria of the primary outcome (60/170, 35% in the corticosteroid group and 106/343, 31% in the noncorticosteroid group). At multivariable analysis corticosteroid treatment was not associated with lower 30-day mortality rate (adjusted odds ratio, 0.59; 95% confidence interval (CI), 0.20–1.74; p 0.33). After inverse probability of treatment weighting, corticosteroids were not associated with lower 30-day mortality (average treatment effect, 0.05; 95% CI, −0.02 to 0.09; p 0.12). However, subgroup analysis revealed that in patients with PO2/FiO2 < 200 mm Hg at admission (135 patients, 52 (38%) treated with corticosteroids), corticosteroid treatment was associated with a lower risk of 30-day mortality (23/52, 44% vs. 45/83, 54%; adjusted odds ratio, 0.20; 95% CI, 0.04–0.90; p 0.036). Conclusions: The effect of corticosteroid treatment on mortality might be l

In recent years, the exponential growth in the number of items and products handled by e-commerce sites has led to the introduction of intelligent systems aimed at supporting users during the decision-making proces. Making the choice of a product among thousands of items becomes complicated for consumers, and in response to this problem, recommender systems (RS) are born. These systems are a set of algorithms based on the concept of information filtering and make it possible to reduce the cognitive effort required of users. In this paper we present a model-based RS, belonging to the collaborative filtering (CF) category, for the e-commerce website of the company Nathan Instruments (NI). Thus, the main objective of this paper is to provide an intelligent approach for recommending configurations of hardware components for Computers. This configurator uses clustering algorithms to address the problems associated with small dataset sizes. Finally, in the experimentation and conclusion sections it is reported how the proposed model simplifies the decision process related to the required computer customization in terms of hardware and software components.

In the latest period as the pandemic has spread worldwide, the need to provide students and learners with effective and personalized didactical resources in order to maintain and even empower the potential level of training has significantly increased: specifically, the major necessity is related to the high potentialities in adopting educational models and methods based on artificial intelligence to reduce the cognitive effort made by teachers when searching, interacting with peers, creating and organizing educational resources. Therefore, nowadays there is a greater diffusion of Artificial Intelligence (AI) techniques and methods for educational purposed and platforms. Every training institution, school and organization definitively requires both advanced, smart and user-friendly technological tools to foster education and training process in the present historical situation. AI algorithms have proved to be crucial in different and heterogeneous research domains: among these methods, recommendation systems have been studied and applied in the e-learning domain. Nevertheless, AI and deep learning (DL) models suffer from trust issues, which poses specific problems which are different than in other kinds of technologies. Thus, if similar concepts are not addressed, there is the concrete to obtain models that perform well but exhibit inconsistent behaviors causing confusion during the interaction process between user and system: such confusion can even be a reason for abandonment of AI technology. Among these intelligent algorithms, content filtering mechanisms such as recommendation systems allow to provide explicit and interactive mechanisms to foster effective cooperation between the users and the system. In this paper, we describe the current human interaction level within the recommendation system (RS) for filtering educational resources of the educational platform ”WhoTeach”, which aims to create and personalize educational courses. In particular, we report the

This study deals with the problem of the recognition of dermatological and exanthematic diseases through the use of deep learning techniques in order to diagnose malignant diseases at an early stage and in general to bring the pathology identified by the models to the attention of the person. A fundamental part of the research was the study of the methodologies present in the state of the art and for this reason in this paper we report the studies considered as most relevant. In this paper, two different types of models are reported, the Convolutional Network Disease (CND) model and the CND-InceptionV3 model using the transfer learning technique. The use of these two models made it possible to carry out an experimental phase in which the performance that can be achieved using the ISIC-archive dataset was analysed. Subsequently, the description of the work carried out for the improvement of the dataset through the association of syntactic-semantic information is reported. Finally, in the last section, conclusions are drawn on the values obtained and future developments that can be made to improve the performance of the models reported are reported.

This study stems from the observation of the growing demand, interest and incredible potential offered by Artificial Intelligence and Machine Learning techniques in the fields of content website or ecommerce. In particular, we will study the Content Management System (CMS), a system capable of creating and maintaining websites, forums and applications for customers. These systems have an intrinsic predisposition to provide services based on Machine Learning (ML), thus being able to add to the solutions offered by the most innovative ML systems. These include Sentiment Analysis, the Recommender System and the more innovative Chatbot . Integrating them into a website or web platform very often leads to significantly improved results for the user experience offered to the user. This study aims to analyse the new artificial intelligence techniques and how to implement them in a company. Reporting also a case of Study based on a real agency

Learning and training processes are starting to be affected by the diffusion of Artificial Intelligence (AI) techniques and methods. AI can be variously exploited for supporting education, though especially deep learning (DL) models are normally suffering from some degree of opacity and lack of interpretability. Explainable AI (XAI) is aimed at creating a set of new AI techniques able to improve their output or decisions with more transparency and interpretability. Deep attentional mechanisms proved to be particularly effective for identifying relevant communities and relationships in any given input network that can be exploited with the aim of improving useful information to interpret the suggested decision process. In this paper we provide the first stages of our ongoing research project, aimed at significantly empowering the recommender system of the educational platform”WhoTeach” by means of explainability, to help teachers or experts to create and manage high-quality courses for personalized learning. The presented model is actually our first tentative to start to include explainability in the system. As shown, the model has strong potentialities to provide relevant recommendations. Moreover, it allows the possibility to implement effective techniques to completely reach explainability3

Despite the recent advances in Natural Language Processing (NLP) techniques many issues and inefficiencies arise when it comes to creating a system capable of interacting with the users by means of text conversations. The current techniques rely on the development of chatbots that, however, require designing the conversation flow, defining training questions and associating the expected responses. Even though this process allows the creation of effective question-answering systems, this methodology is not scalable, especially when the answers are to be found in documents. Other approaches, instead, rely on graph embedding techniques and graph neural networks to define the best answer given a question. These methods, however, require set-up training routines and to dispose of the ground truth that, in general, is difficult to retrieve or create for real industrial applications. In this paper we introduce a conversational framework for semantic question answering. Our work relies on knowledge graphs and the use of machine learning for determining the best answer given a question associated with the content of the knowledge graph. In addition, by leveraging text mining techniques we are able to identify the best set of answers that suit the question that are further filtered by means of deep learning algorithms.