Pasar al contenido principal

AGENTES INTELIGENTES PARA ASESORAR EN PRIVACIDAD EN REDES SOCIALES

TIN2017-89156-R

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal de I+D+i Orientado a los Retos de la Sociedad
Convocatoria Proyectos I+D+i Retos de Investigación
Año convocatoria 2017
Plan Estatal 2017-2020 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSITAT POLITÈCNICA DE VALÈNCIA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 31
Encontrada(s) 1 página(s)

RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Tie and tag: A study of tie strength and tags for photo sharing

  • López Fogués, Ricard
  • Such Aparicio, José Miguel|||0000-0002-6041-178X
  • Espinosa Minguet, Agustín Rafael|||0000-0001-8070-3344
  • García-Fornes, A|||0000-0003-4482-8793
[EN] Tie strength and tags have been separately suggested as possible attributes for photo access controls in Social Network Services. However, an evaluation is missing about the benefits/drawbacks of adding one or both of these attributes to the ones already used in access controls for Social Network Services (groups and individuals). In this paper, we describe an experiment with 48 participants using access controls that include tie strength and tags (separately and simultaneously) together with attributes for groups and individuals. We analyze the results using several quantitative and qualitative metrics. We find that users consider these two new attributes useful in defining their sharing policies, and they prefer to employ access controls that consider tags and tie strength jointly. Specifically, users believe that tie strength improves policy understandability and that tags help them define sharing policies faster. However, we also observe that when users employ these two attributes they tend to make more mistakes in terms of the resulting sharing policy. We hypothesize that this could be caused by the lack of experience using tie strength and tags in access controls., This research was supported by the Ministerio de Economia y Competitividad (http://www.mineco.gob.es/), grant number: TIN2014-55206-R and TIN2017-89156-R to A.G.F. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Towards a Persuasive Recommender for Bike Sharing Systems: A Defeasible Argumentation Approach

  • Diez-Alba, Carlos|||0000-0001-5055-3525
  • Palanca Cámara, Javier|||0000-0002-6209-9603
  • Sanchez-Anguix, Víctor|||0000-0003-4851-0037
  • Heras, Stella|||0000-0001-6212-9377
  • Giret Boggino, Adriana Susana|||0000-0002-2311-0785
  • Julian, Vicente|||0000-0002-2743-6037
[EN] This work proposes a persuasion model based on argumentation theory and users' characteristics for improving the use of resources in bike sharing systems, fostering the use of the bicycles and thus contributing to greater energy sustainability by reducing the use of carbon-based fuels. More specifically, it aims to achieve a balanced network of pick-up and drop-off stations in urban areas with the help of the users, thus reducing the dedicated management trucks that redistribute bikes among stations. The proposal aims to persuade users to choose different routes from the shortest route between a start and an end location. This persuasion is carried out when it is not possible to park the bike in the desired station due to the lack of parking slots, or when the user is highly influenceable. Differently to other works, instead of employing a single criteria to recommend alternative stations, the proposed system can incorporate a variety of criteria. This result is achieved by providing a defeasible logic-based persuasion engine that is capable of aggregating the results from multiple recommendation rules. The proposed framework is showcased with an example scenario of a bike sharing system., This work was supported by the projects TIN2015-65515-C4-1-R and TIN2017-89156-R of the Spanish government, and by the grant program for the recruitment of doctors for the Spanish system of science and technology (PAID-10-14) of the Universitat Politècnica de València.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Using Keystroke Dynamics in a Multi-Agent System for User Guiding in Online Social Networks

  • Aguado-Sarrió, Guillem
  • Julian, Vicente|||0000-0002-2743-6037
  • García-Fornes, A|||0000-0003-4482-8793
  • Espinosa Minguet, Agustín Rafael|||0000-0001-8070-3344
[EN] Nowadays there is a strong integration of online social platforms and applications with our daily life. Such interactions can make risks arise and compromise the information we share, thereby leading to privacy issues. In this work, a proposal that makes use of a software agent that performs sentiment analysis and another performing stress analysis on keystroke dynamics data has been designed and implemented. The proposal consists of a set of new agents that have been integrated into a multi-agent system (MAS) for guiding users interacting in online social environments, which has agents for sentiment and stress analysis on text. We propose a combined analysis using the different agents. The MAS analyzes the states of the users when they are interacting, and warns them if the messages they write are deemed negative. In this way, we aim to prevent potential negative outcomes on social network sites (SNSs). We performed experiments in the laboratory with our private SNS Pesedia over a period of one month, so we gathered data about text messages and keystroke dynamics data, and used the datasets to train the artificial neural networks (ANNs) of the agents. A set of experiments was performed for discovering which analysis is able to detect a state of the user that propagates more in the SNS, so it may be more informative for the MAS. Our study will help develop future intelligent systems that utilize user data in online social environments for guiding or helping them in their social experience., This work was funded by the project TIN2017-89156-R of the Spanish government.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Using a Hybrid Recommending System for Learning Videos in Flipped Classrooms and MOOCs

  • Jordán, Jaume|||0000-0003-0400-9136
  • Valero Cubas, Soledad|||0000-0003-4565-326X
  • Turró Ribalta, Carlos|||0000-0002-3840-9405
  • Botti V.|||0000-0002-6507-2756
[EN] New challenges in education require new ways of education. Higher education has adapted to these new challenges by means of offering new types of training like massive online open courses and by updating their teaching methodology using novel approaches as flipped classrooms. These types of training have enabled universities to better adapt to the challenges posed by the pandemic. In addition, high quality learning objects are necessary for these new forms of education to be successful, with learning videos being the most common learning objects to provide theoretical concepts. This paper describes a new approach of a previously presented hybrid learning recommender system based on content-based techniques, which was capable of recommend useful videos to learners and lecturers from a learning video repository. In this new approach, the content-based techniques are also combined with a collaborative filtering module, which increases the probability of recommending relevant videos. This hybrid technique has been successfully applied to a real scenario in the central video repository of the Universitat Politècnica de València., This research was partially supported by MINECO/FEDER RTI2018-095390-B-C31 and TIN2017-89156-R projects of the Spanish government, and PROMETEO/2018/002 project of Generalitat Valenciana.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Comparison of Predictive Models with Balanced Classes Using the SMOTE Method for the Forecast of Student Dropout in Higher Education

  • Flores, Vaneza
  • Heras, Stella|||0000-0001-6212-9377
  • Julian, Vicente|||0000-0002-2743-6037
[EN] Based on the premise that university student dropout is a social problem in the university ecosystem of any country, technological leverage is a way that allows us to build technological proposals to solve a poorly met need in university education systems. Under this scenario, the study presents and analyzes eight predictive models to forecast university dropout, based on data mining methods and techniques, using WEKA for its implementation, with a dataset of 4365 academic records of students from the National University of Moquegua (UNAM), Peru. The objective is to determine which model presents the best performance indicators to forecast and prevent student dropout. The study aims to propose and compare the accuracy of eight predictive models with balanced classes, using the SMOTE method for the generation of synthetic data. The results allow us to confirm that the predictive model based on Random Forest is the one that presents the highest accuracy and robustness. This study is of great interest to the educational community as it allows for predicting the possible dropout of a student from a university career and being able to take corrective actions both at a global and individual level. The results obtained are highly interesting for the university in which the study has been carried out, obtaining results that generally outperform the results obtained in related works., This work was partially supported by the Spanish Government project TIN2017-89156-R, and the Valencian Government project PROMETEO/2018/002. The research was developed thanks to the support of the National University of Moquegua, which provided the information for the creation of the dataset.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Contribución de congreso (ConferenceOutput)

Recommending Learning Videos for MOOCs and Flipped Classrooms

  • Jordán, Jaume|||0000-0003-0400-9136
  • Valero Cubas, Soledad|||0000-0003-4565-326X
  • Turró Ribalta, Carlos|||0000-0002-3840-9405
  • Botti V.|||0000-0002-6507-2756
[EN] New teaching approaches are emerging in higher education, such as flipped classrooms. In addition, academic institutions are offering new types of training like Massive Online Open Courses. Both of these new ways of education require high-quality learning objects for their success, with learning videos being the most common to provide theoretical concepts. This paper describes a hybrid learning recommender system based on content-based techniques, which is able to recommend useful videos to learners and teachers from a learning video repository. This hybrid technique has been successfully applied to a real scenario such as the central video repository of the Universitat Politècnica de València., This work was partially supported by MINECO/FEDER RTI2018-095390-B-C31 and TIN2017-89156-R projects of the Spanish government, and PROMETEO/2018/002 project of Generalitat Valenciana. J. Jordán and V. Botti are funded by UPV PAID-06-18 project. J. Jordán is also funded by grant APOSTD/2018/010 of Generalitat Valenciana - Fondo Social Europeo.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Enhancing the privacy risk awareness of teenagers in online social networks through soft-paternalism mechanisms

  • Alemany-Bordera, José
  • Del Val Noguera, Elena|||0000-0002-1279-3429
  • Alberola Oltra, Juan Miguel|||0000-0002-5486-5638
  • García-Fornes, A|||0000-0003-4482-8793
[EN] Privacy Risk in Online Social Networks (OSNs) is one of the main concerns that has increased in the last few years. Even though social network applications provide mechanisms to control risk, teenagers are not often aware of the privacy risks of disclosing information in online social networks. The privacy decision-making process is complex and users often do not have full knowledge and enough time to evaluate all potential scenarios. They do not consider the audience that will have access to disclosed information or the risk if the information continues to spread and reaches an unexpected audience. To deal with these issues, we propose two soft-paternalism mechanisms that provide information to the user about the privacy risk of publishing information on a social network. That privacy risk is based on a complex privacy metric. To evaluate the mechanisms, we performed an experiment with 42 teenagers. The proposed mechanisms were included in a social network called Pesedia. The results show that there are significant differences in teenagers¿ behaviors towards better privacy practices when the mechanisms are included in the network., This work is partially supported by the Spanish Government project TIN2017-89156-R and the FPI grant BES-2015-074498.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Contribución de congreso (ConferenceOutput)

Automatic Detection of Sensitive Information in Educative Social Networks

  • Botti-Cebriá, Víctor
  • Del Val Noguera, Elena|||0000-0002-1279-3429
  • García-Fornes, A|||0000-0003-4482-8793
[EN] Detecting sensitive information with privacy in mind is a relevant issue on Social Networks. It is often difficult for users to manage the privacy associated with their posts on social networks taking into account their possible consequences. The main objective of this work is to provide users information about the sensitivity of the information they will share when they decide to publish a message in online media. For this purpose, an assistant agent to detect sensitive information based on different types of categories detected in the message (i.e., location, personal data, health, personal attacks, emotions, etc.) is proposed. Entity recognition libraries, ontologies, dictionaries, and sentiment analysis will be used to detect the different categories. This agent is integrated into the social network Pesedia, aimed for children and teenagers, and through a soft-paternalism mechanism provides information to users about the sensitivity of certain content and help them in making decisions about its publication. The agent decision process will be evaluated with a dataset elaborated from messages of the social network Twitter., This work is supported by the Spanish Government project TIN2017-89156-R.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Transformer-Based Models for Automatic Identification of Argument Relations: A Cross-Domain Evaluation

  • Ruiz-Dolz, Ramon
  • Alemany-Bordera, José
  • Heras, Stella|||0000-0001-6212-9377
  • García-Fornes, A|||0000-0003-4482-8793
[EN] Argument mining is defined as the task of automatically identifying and extracting argumentative components (e.g., premises, claims, etc.) and detecting the existing relations among them (i.e., support, attack, rephrase, no relation). One of the main issues when approaching this problem is the lack of data, and the size of the publicly available corpora. In this work, we use the recently annotated US2016 debate corpus. US2016 is the largest existing argument annotated corpus, which allows exploring the benefits of the most recent advances in natural language processing in a complex domain like argument (relation) mining. We present an exhaustive analysis of the behavior of transformer-based models (i.e., BERT, XLNET, RoBERTa, DistilBERT, and ALBERT) when predicting argument relations. Finally, we evaluate the models in five different domains, with the objective of finding the less domain-dependent model. We obtain a macro F1-score of 0.70 with the US2016 evaluation corpus, and a macro F1-score of 0.61 with the Moral Maze cross-domain corpus., This work was supported in part by the Spanish Government project under Grant TIN2017-89,156-R, in part by the FPI under grant BES-2015-074,498, and in part by the Valencian Government project under Grant PROMETEO/2018/002. The authors gratefully acknowledge the support of NVIDIA Corporation with the donation of the Titan V GPUs used for this research.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

A Multi-Agent System for guiding users in on-line social environments

  • Aguado-Sarrió, Guillem
  • Julian, Vicente|||0000-0002-2743-6037
  • García-Fornes, A|||0000-0003-4482-8793
  • Espinosa Minguet, Agustín Rafael|||0000-0001-8070-3344
[EN] The present work is a study of the detection of negative affective or emotional states, the high-stress levels that people have using social network sites (SNSs), and the effect that this negative state or stress level has on the repercussions of posted messages. We aim to discover to what extent a user that has a state detected as negative by an analyzer (Sentiment analyzer and Stress analyzer) can affect other users and generate negative repercussions, and also determine whether it is more suitable to predict a future negative situation using different analyzers. We propose two different methods for creating a combined model of sentiment and stress, and we use them in our experimentation to discern which one is more suitable for predicting future negative situations that could arise from the interaction between users, and in what context. Additionally, we designed a Multi-Agent System (MAS) that integrates the analyzers to protect or advise users on a SNS. We have conducted this study to help build future systems that prevent negative situations where a user that has a negative state creates a repercussion in the SNS. This can help users avoid getting into a bad mood or help avoid privacy issues (e.g. a user that has a negative state posting information that the user does not really want to post)., This work was supported by the project TIN2017-89156-R of the Ministry of Economy, Industry and Competitiveness, Government of Spain.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

A CBR for integrating sentiment and stress analysis for guiding users on social network sites

  • Aguado, Guillem
  • Julian, Vicente|||0000-0002-2743-6037
  • García-Fornes, A|||0000-0003-4482-8793
  • Espinosa Minguet, Agustín Rafael|||0000-0001-8070-3344
[EN] This work presents a Case-Based Reasoning (CBR) module that integrates sentiment and stress analysis on text and keystroke dynamics data with context information of users interacting on Social Network Sites (SNSs). The context information used in this work is the history of positive or negative messages of the user, and the topics being discussed on the SNSs. The CBR module uses this data to generate useful feedback for users, providing them with warnings if it detects potential future negative repercussions caused by the interaction of the users in the system. We aim to help create a safer and more satisfactory experience for users on SNSs or in other social environments. In a set of experiments, we compare the effectiveness of the CBR module to the effectiveness of different affective state detection methods. We compare the capacity to detect cases of messages that would generate future problems or negative repercussions on the SNS. For this purpose, we use messages generated in a private SNS, called Pesedia. In the experiments in the laboratory, the CBR module managed to outperform the other proposed analyzers in almost every case. The CBR module was fine-tuned to explore its performance when populating the case base with different configurations., This work was financed by the project TIN2017-89156-R of the Ministry of Economy, Industry and Competitiveness, Government of Spain.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Contribución de congreso (ConferenceOutput)

Simulating Users in a Social Media Platform Using Multi-agent Systems

  • Pérez-García, Daniel
  • Argente, Estefanía|||0000-0002-5431-3868
[EN] The massive use of social media makes it increasingly easy to find highly sensitive information about almost anyone on the Internet. Despite the efforts of social media platforms to provide their users with tools to manage their privacy, these are proving insufficient due to their complexity. For this reason, it has been considered necessary to develop a software tool based on a multi-agent system to help users to improve and correct their bad behavior by using automation mechanisms and transmitting the information in a natural way for them, replicating the behavior of a human being. The aim of our work is to implement a multi- agent system where agents interact organically with each other and with human users on the PESEDIA social network, so that they can support user in a non-intrusive manner, using paternalistic techniques through actions available on the social network., This work has been funded thanks to the Spanish Government through project TIN2017-89156-R and predoctoral contract PRE2018-084940.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

A Review of Privacy Decision-making Mechanisms in Online Social Networks

  • Alemany, J.
  • Del Val Noguera, Elena|||0000-0002-1279-3429
  • García-Fornes, A|||0000-0003-4482-8793
[EN] Personal information of online social networks (OSNs) is governed by the privacy policies chosen by users besides OSN's policies. Users make these decisions using privacy mechanisms, but privacy problems and regrets are daily reported. This article reviews current privacy mechanisms and solutions. For this, we analyze all the sub-decisions and elements of online communication involved in the privacy decision-making process. However, the differences in users' motivations and the disclosure of too sensitive information (among others) can lead to loss of privacy. In this work, we identify requirements such as automation, preference-centered, relationship-based, and multi-party privacy mechanisms, which have been more researched. But also other requirements (recently emerged), such as privacy preservation with risk metrics, explainability, and ephemeral messages. We explore all the advances made in the literature, and we have seen that most of these have been focused on matching the users' preferences with their decision (which is not appropriate, because users cannot evaluate all of the potential privacy scenarios) instead of assessing privacy risk metrics, adaptation, and explainability. Therefore, we have identified open challenges, such as metrics for assessing privacy risks, explainable solutions for users, ephemeral communication solutions, and the application of these requirements to the multi-party privacy scenario., This work is partially supported by the Spanish Government Project No. TIN2017-89156-R and the FPI Grant No. BES-2015-074498.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

A genetic algorithm for group formation in elderly communities

  • Alberola Oltra, Juan Miguel|||0000-0002-5486-5638
  • Del Val Noguera, Elena|||0000-0002-1279-3429
  • Julian, Vicente|||0000-0002-2743-6037
  • Araujo, Angelo
  • Novais, Paulo
[EN] Over the last few years, agent technology has proposed different approaches for the development of systems that support the formation and dynamic management of teams or groups. Following this idea, this paper proposes a model for group formation in elderly communities defining it as a Coalition Structure Generation Problem. Specifically, the proposal has been used to analyse which is the best way to organize older people into activity groups in elderly communities. The model parameters are physical requirements, preferences and social relationships, being the model able to learn from each execution and improve the future configurations. The results show near-optimal solutions to all proposed scenarios, beating greatly the computational time of CPLEX, and flexibility to deal with different scenarios and configurations., This work has been supported by COMPETE: POCI-01-0145-FEDER-007043 and FCT - Fundacao para a Ciencia e Tecnologia within the Project Scope: UID/CEC/00319/2013. A. Costa thanks the FCT - Fundacao para a Ciencia e Tecnologia the Post-Doc scholarship with the Ref. SFRH/BPD/102696/2014. This work is also partially supported by the MINECO/FEDER TIN2017-89156-R and the Post-Doc scholarship with the Ref. SP20170057.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

A fuzzy appraisal model for affective agents adapted to cultural environments using the pleasure and arousal dimensions

  • Taverner-Aparicio, Joaquín José|||0000-0002-5163-5335
  • Vivancos, Emilio|||0000-0002-0213-0234
  • Botti V.|||0000-0002-6507-2756
[EN] Humans use rather vague and imprecise words to express emotions. Therefore, fuzzy logic allows computational affective models to use emotions in the same way that human beings express them. However, people from different cultures and languages assign different meanings to the same emotion word. Unfortunately, there are still no affective computing models that really take these two factors into consideration. In this paper, we propose a fuzzy model of appraisal for multi-agent systems that is adapted to Spanish-speakers. Our methodology has two steps. First, the agent evaluates an event using a set of fuzzy appraisal rules that returns a fuzzy emotion. Then, a defuzzification process returns the Pleasure and Arousal dimensions of the emotion that will be internally represented as a vector in a two-dimensional space. This vector is used to update the agent's mood according to the agent's personality. The agent can express this internal emotional state using a fuzzification process that translates the agent's mood into a fuzzy emotion. This fuzzification process uses the results of an experiment to generate a fuzzy emotion that is adapted to the cultural environment in which the agent is located. This methodology can be easily adapted to other languages., This work is partially supported by the Spanish Government project TIN2017-89156-R, Generalitat Valenciana and European Social Fund by the FPI grant ACIF/2017/085, and GVA-CEICE project PROMETEO/2018/002.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Estimation of privacy risk through centrality metrics

  • Alemany-Bordera, José
  • Del Val Noguera, Elena|||0000-0002-1279-3429
  • Alberola Oltra, Juan Miguel|||0000-0002-5486-5638
  • García-Fornes, A|||0000-0003-4482-8793
[EN] Users are not often aware of privacy risks and disclose information in online social networks. They do not consider the audience that will have access to it or the risk that the information continues to spread and may reach an unexpected audience. Moreover, not all users have the same perception of risk. To overcome these issues, we propose a Privacy Risk Score (PRS) that: (1) estimates the reachability of an user¿s sharing action based on the distance between the user and the potential audience; (2) is described in levels to adjust to the risk perception of individuals; (3) does not require the explicit interaction of individuals since it considers information flows; and (4) can be approximated by centrality metrics for scenarios where there is no access to data about information flows. In this case, if there is access to the network structure, the results show that global metrics such as closeness have a high degree of correlation with PRS. Otherwise, local and social centrality metrics based on ego-networks provide a suitable approximation to PRS. The results in real social networks confirm that local and social centrality metrics based on degree perform well in estimating the privacy risk of users., This work is partially supported by the Spanish Government project TIN2014-55206-R and FPI grant BES-2015-074498.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Detection and nudge-intervention on sensitive information in social networks

  • Alemany, J.
  • Botti-Cebriá, V.
  • Del Val Noguera, Elena|||0000-0002-1279-3429
  • García-Fornes, A|||0000-0003-4482-8793
[EN] Detecting sensitive information considering privacy is a relevant issue on Online Social Networks (OSNs). It is often difficult for users to manage the privacy associated with their posts on social networks taking into account all the possible consequences. The aim of this work is to provide information about the sensitivity of the content of a publication when a user is going to share it in OSN. For this purpose, we developed a privacy-assistant agent that detects sensitive information. Based on this information, the agent provides a message through a nudge mechanism warning about the possible risks of sharing the message. To avoid being annoying, the agent also considers the user's previous behaviour (e.g. if he previously ignored certain nudges) and adapts the messages it sends to give more relevance to those categories that are more important to the user from the point of view of the privacy risk. This agent was integrated into the social network Pesedia. We analysed the performance of different models to detect a set of sensitive categories (i.e. location, medical, drug/alcohol, emotion, personal attacks, stereotyping, family and association details, personal details and personally identifiable information) in a dataset of tweets in Spanish. The model that obtained the best results (i.e. F1 and accuracy) and that was finally integrated into the privacy-assistant agent was transformer-based., This work is supported by the Spanish Government project TIN2017-89156-R.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

A multidimensional culturally adapted representation of emotions for affective computational simulation and recognition

  • Taverner-Aparicio, Joaquín José|||0000-0002-5163-5335
  • Vivancos, Emilio|||0000-0002-0213-0234
  • Botti V.|||0000-0002-6507-2756
[EN] One of the main challenges in affective computing is the development of models to represent the information that is inherent to emotions. It is necessary to consider that the terms used by humans to name emotions depend on the culture and language used. This article presents an experiment-based method to represent and adapt emotion terms to different cultural environments. We propose using circular boxplots to analyze the distribution of emotions in the Pleasure-Arousal space. From the results of this analysis, we define a new cross-cultural representation model of emotions in which each emotion term is assigned to an area in the Pleasure-Arousal space. An emotion is represented by a vector in which the direction indicates the type, and the module indicates the intensity of the emotion. We propose two methods based on fuzzy logic to represent and express emotions: the emotion representation process in which the term associated with the recognized emotion is defuzzified and projected as a vector in the Pleasure-Arousal space; and the emotion expression process in which a fuzzification of the vector is produced, generating a fuzzy emotion term that is adapted to the culture and language in which the emotion will be used., This work was supported in part by the Spanish Government project TIN2017-89156-R, Generalitat Valenciana, and European Social Fund by the FPI Grant ACIF/2017/085, in part GVA-CEICE project PROMETEO/2018/002, and TAILOR, a project funded by EU Horizon 2020 research and innovation programme under GA No 952215.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Towards Aiding Decision-Making in Social Networks by Using Sentiment and Stress Combined Analysis

  • Aguado-Sarrió, Guillem
  • Julian, Vicente|||0000-0002-2743-6037
  • García-Fornes, A|||0000-0003-4482-8793
[EN] The present work is a study of the detection of negative emotional states that people have using social network sites (SNSs), and the effect that this negative state has on the repercussions of posted messages. We aim to discover in which grade a user having an affective state considered negative by an Analyzer can affect other users and generate bad repercussions. Those Analyzers that we propose are a Sentiment Analyzer, a Stress Analyzer and a novel combined Analyzer. We also want to discover what Analyzer is more suitable to predict a bad future situation, and in what context. We designed a Multi-Agent System (MAS) that uses different Analyzers to protect or advise users. This MAS uses the trained and tested Analyzers to predict future bad situations in social media, which could be triggered by the actions of a user that has an emotional state considered negative. We conducted an experimentation with different datasets of text messages from Twitter.com to examine the ability of the system to predict bad repercussions, by comparing the polarity, stress level or combined value classification of the messages that are replies to the ones of the messages that originated them., This work was supported by the project TIN2017-89156-R of the Spanish government.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Toward the Prevention of Privacy Threats: How Can We Persuade Our Social Network Platform Users?

  • Ruiz-Dolz, Ramon
  • Alemany, José
  • Heras, Stella|||0000-0001-6212-9377
  • García-Fornes, A|||0000-0003-4482-8793
[EN] Complex decision-making problems, such as the privacy policy selection, when sharing content in online social network (OSN) platforms can significantly benefit from artificial intelligence systems. With the use of computational argumentation, it is possible to persuade human users to modify their initial decisions to avoid potential privacy threats and violations. In this paper, we present a study performed with the participation of 186 teenage users aimed at analyzing their behaviors when we try to persuade them to modify the post/publication of sensitive content on OSN platforms with different arguments. The results of the study revealed that the personality traits and the social interaction data (e.g., number of comment posts, friends, and likes) of our participants were significantly correlated with the persuasive power of the arguments. Therefore, these sets of features can be used to model OSN users and estimate the persuasive power of different arguments when used in human-computer interactions. The findings presented in this paper are helpful for personalizing decision support systems aimed at educating and preventing privacy violations on OSN platforms using arguments., This work was partially supported by the Spanish Government (Project TIN2017-89156-R, PID2020- 113416RB-I00), and the Valencian Government (Project PROMETEO/2018/002).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Normative Emotional Agents: a viewpoint paper

  • Argente, Estefanía|||0000-0002-5431-3868
  • Del Val Noguera, Elena|||0000-0002-1279-3429
  • Botti V.|||0000-0002-6507-2756
  • Pérez-García, Daniel
[EN] Human social relationships imply conforming to the norms, behaviors and cultural values of the society, but also socialization of emotions, to learn how to interpret and show them. In multiagent systems, much progress has been made in the analysis and interpretation of both emotions and norms. Nonetheless, the relationship between emotions and norms has hardly been considered and most normative agents do not consider emotions, or vice-versa. In this article, we provide an overview of relevant aspects within the area of normative agents and emotional agents. First we focus on the concept of norm, the different types of norms, its life cycle and a review of multiagent normative systems. Secondly, we present the most relevant theories of emotions, the life cycle of an agent¿s emotions, and how emotions have been included through computational models in multiagent systems. Next, we present an analysis of proposals that integrate emotions and norms in multiagent systems. From this analysis, four relationships are detected between norms and emotions, which we analyze in detail and discuss how these relationships have been tackled in the reviewed proposals. Finally, we present a proposal for an abstract architecture of a Normative Emotional Agent that covers these four norm-emotion relationships., This work was supported by the Spanish Government project TIN2017-89156- R, the Generalitat Valenciana project PROMETEO/2018/002 and the Spanish Goverment PhD Grant PRE2018-084940.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Contribución de congreso (ConferenceOutput)

Towards a Classifier to Recognize Emotions Using Voice to Improve Recommendations

  • Fuentes-López, José Manuel
  • Rincón Arango, Jaime Andrés
  • Taverner-Aparicio, Joaquín José|||0000-0002-5163-5335
  • Botti V.|||0000-0002-6507-2756
[EN] The recognition of emotions in tone voice is currently a tool with a high potential when it comes to making recommendations, since it allows to personalize recommendations using the mood of the users as information. However, recognizing emotions using tone of voice is a complex task since it is necessary to pre-process the signal and subsequently recognize the emotion. Most of the current proposals use recurrent networks based on sequences with a temporal relationship. The disadvantage of these networks is that they have a high runtime, which makes it difficult to use in real-time applications. On the other hand, when defining this type of classifier, culture and language must be taken into account, since the tone of voice for the same emotion can vary depending on these cultural factors. In this work we propose a culturally adapted model for recognizing emotions from the voice tone using convolutional neural networks. This type of network has a relatively short execution time allowing its use in real time applications. The results we have obtained improve the current state of the art, reaching 93.6% success over the validation set., This work is partially supported by the Spanish Government project TIN2017-89156-R, GVA-CEICE project PROMETEO/2018/002, Generalitat Valenciana and European Social Fund FPI grant ACIF/2017/085, Universitat Politecnica de Valencia research grant (PAID-10-19), and by the Spanish Government (RTI2018-095390-B-C31).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Tesis doctoral (DoctoralThesis)

Computational Argumentation for the Automatic Analysis of Argumentative Discourse and Human Persuasion

  • Ruiz Dolz, Ramon
[ES] La argumentación computacional es el área de investigación que estudia y analiza el uso de distintas técnicas y algoritmos que aproximan el razonamiento argumentativo humano desde un punto de vista computacional. En esta tesis doctoral se estudia el uso de distintas técnicas propuestas bajo el marco de la argumentación computacional para realizar un análisis automático del discurso argumentativo, y para desarrollar técnicas de persuasión computacional basadas en argumentos. Con estos objetivos, en primer lugar se presenta una completa revisión del estado del arte y se propone una clasificación de los trabajos existentes en el área de la argumentación computacional. Esta revisión nos permite contextualizar y entender la investigación previa de forma más clara desde la perspectiva humana del razonamiento argumentativo, así como identificar las principales limitaciones y futuras tendencias de la investigación realizada en argumentación computacional. En segundo lugar, con el objetivo de solucionar algunas de estas limitaciones, se ha creado y descrito un nuevo conjunto de datos que permite abordar nuevos retos y investigar problemas previamente inabordables (e.g., evaluación automática de debates orales). Conjuntamente con estos datos, se propone un nuevo sistema para la extracción automática de argumentos y se realiza el análisis comparativo de distintas técnicas para esta misma tarea. Además, se propone un nuevo algoritmo para la evaluación automática de debates argumentativos y se prueba con debates humanos reales. Finalmente, en tercer lugar se presentan una serie de estudios y propuestas para mejorar la capacidad persuasiva de sistemas de argumentación computacionales en la interacción con usuarios humanos. De esta forma, en esta tesis se presentan avances en cada una de las partes principales del proceso de argumentación computacional (i.e., extracción automática de argumentos, representación del conocimiento y razonamiento basados en argumentos, e interacción humano-computador basada en argumentos), así como se proponen algunos de los cimientos esenciales para el análisis automático completo de discursos argumentativos en lenguaje natural., [CA] L'argumentació computacional és l'àrea de recerca que estudia i analitza l'ús de distintes tècniques i algoritmes que aproximen el raonament argumentatiu humà des d'un punt de vista computacional. En aquesta tesi doctoral s'estudia l'ús de distintes tècniques proposades sota el marc de l'argumentació computacional per a realitzar una anàlisi automàtic del discurs argumentatiu, i per a desenvolupar tècniques de persuasió computacional basades en arguments. Amb aquestos objectius, en primer lloc es presenta una completa revisió de l'estat de l'art i es proposa una classificació dels treballs existents en l'àrea de l'argumentació computacional. Aquesta revisió permet contextualitzar i entendre la investigació previa de forma més clara des de la perspectiva humana del raonament argumentatiu, així com identificar les principals limitacions i futures tendències de la investigació realitzada en argumentació computacional. En segon lloc, amb l'objectiu de sol$\cdot$lucionar algunes d'aquestes limitacions, hem creat i descrit un nou conjunt de dades que ens permet abordar nous reptes i investigar problemes prèviament inabordables (e.g., avaluació automàtica de debats orals). Conjuntament amb aquestes dades, es proposa un nou sistema per a l'extracció d'arguments i es realitza l'anàlisi comparativa de distintes tècniques per a aquesta mateixa tasca. A més a més, es proposa un nou algoritme per a l'avaluació automàtica de debats argumentatius i es prova amb debats humans reals. Finalment, en tercer lloc es presenten una sèrie d'estudis i propostes per a millorar la capacitat persuasiva de sistemes d'argumentació computacionals en la interacció amb usuaris humans. D'aquesta forma, en aquesta tesi es presenten avanços en cada una de les parts principals del procés d'argumentació computacional (i.e., l'extracció automàtica d'arguments, la representació del coneixement i raonament basats en arguments, i la interacció humà-computador basada en arguments), així com es proposen alguns dels fonaments essencials per a l'anàlisi automàtica completa de discursos argumentatius en llenguatge natural., [EN] Computational argumentation is the area of research that studies and analyses the use of different techniques and algorithms that approximate human argumentative reasoning from a computational viewpoint. In this doctoral thesis we study the use of different techniques proposed under the framework of computational argumentation to perform an automatic analysis of argumentative discourse, and to develop argument-based computational persuasion techniques. With these objectives in mind, we first present a complete review of the state of the art and propose a classification of existing works in the area of computational argumentation. This review allows us to contextualise and understand the previous research more clearly from the human perspective of argumentative reasoning, and to identify the main limitations and future trends of the research done in computational argumentation. Secondly, to overcome some of these limitations, we create and describe a new corpus that allows us to address new challenges and investigate on previously unexplored problems (e.g., automatic evaluation of spoken debates). In conjunction with this data, a new system for argument mining is proposed and a comparative analysis of different techniques for this same task is carried out. In addition, we propose a new algorithm for the automatic evaluation of argumentative debates and we evaluate it with real human debates. Thirdly, a series of studies and proposals are presented to improve the persuasiveness of computational argumentation systems in the interaction with human users. In this way, this thesis presents advances in each of the main parts of the computational argumentation process (i.e., argument mining, argument-based knowledge representation and reasoning, and argument-based human-computer interaction), and proposes some of the essential foundations for the complete automatic analysis of natural language argumentative discourses., This thesis has been partially supported by the Generalitat Valenciana project PROMETEO/2018/002 and by the Spanish Government projects TIN2017-89156-R and PID2020-113416RB-I00
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

A Review on MAS-Based Sentiment and Stress Analysis User-Guiding and Risk-Prevention Systems in Social Network Analysis

  • Aguado-Sarrió, Guillem
  • Julian, Vicente|||0000-0002-2743-6037
  • García-Fornes, A|||0000-0003-4482-8793
  • Espinosa Minguet, Agustín Rafael|||0000-0001-8070-3344
[EN] In the current world we live immersed in online applications, being one of the most present of them Social Network Sites (SNSs), and different issues arise from this interaction. Therefore, there is a need for research that addresses the potential issues born from the increasing user interaction when navigating. For this reason, in this survey we explore works in the line of prevention of risks that can arise from social interaction in online environments, focusing on works using Multi-Agent System (MAS) technologies. For being able to assess what techniques are available for prevention, works in the detection of sentiment polarity and stress levels of users in SNSs will be reviewed. We review with special attention works using MAS technologies for user recommendation and guiding. Through the analysis of previous approaches on detection of the user state and risk prevention in SNSs we elaborate potential future lines of work that might lead to future applications where users can navigate and interact between each other in a more safe way., This work was funded by the project TIN2017-89156-R of the Spanish government.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

&quot, Who should I grant access to my post?&quot, : Identifying the most suitable privacy decisions on online social networks

  • Alemany-Bordera, José
  • Del Val Noguera, Elena|||0000-0002-1279-3429
  • García-Fornes, A|||0000-0003-4482-8793
[EN] Purpose - Online social networks (OSNs) provide users with mechanisms such as social circles and individual selection to define the audiences (i.e., privacy policy) of the shared information. This privacy decision-making process is a hard and tedious task for users because they have to assess the cost-benefit in a complex environment. Moreover, little is known about how users assess the cost-benefit of matching the elements of online communication and their interests. Therefore, the purpose of this paper is to develop and test a research model to understand the impact that the types of receivers and the sensitivity of messages have on privacy decisions. Design/methodology/approach - A study was conducted to understand how users evaluate the cost-benefit of the disclosure action in online social networks for the different types of receivers identified and the sensitivity of the message. Data from 400 respondents was collected and analyzed using partial least squares modeling. Findings - The findings of this study demonstrated a trade-off variance between the perceived cost-benefit and the disclosure of sensitive information with different receiver types. Disclosing personal information with trusted receivers, influencer receivers and receivers from the circle of coworkers had a positive significant effect on social capital building. Conversely, disclosing personal information with receivers from the circle of family or unknown receivers had a significant negative effect on social capital building and even a significant positive effect on privacy concerns. Originality/value - Recent literature has documented the increasing interest of the research community in understanding users' concerns and interests in making the most suitable privacy decisions. However, most researchers have worked on understanding the disclosure action from a user-centered perspective and have not considered all of the elements of online communication. This study puts the focus on all of the elements of communication during disclosure actions, taking into account the properties of the message and receivers and the impact on users' cost benefit value., This work is partially supported by the Spanish Government project TIN2017-89156-R, and the FPI grant BES-2015-074498.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Metrics for privacy assessment when sharing information in online social networks

  • Alemany-Bordera, José
  • Del Val Noguera, Elena|||0000-0002-1279-3429
  • Alberola Oltra, Juan Miguel|||0000-0002-5486-5638
  • García-Fornes, A|||0000-0003-4482-8793
[EN] Privacy risk in Online Social Networks has become an important social concern. Users, with different perceptions of risk, share information without considering the audience that has access to the information disclosed or how far a publication will go. According to this, we propose two metrics (Audience and Reachability) based on information flows and friendship layers that indicate the privacy risk of sharing information, addressing the posts¿ scope and invisible audience. We assess these metrics through agent simulations in well-known models of networks. The findings show a strong relationship between metrics and structural centrality network properties. We also studied scenarios where there is no previous information about users activity or the information about the traces of the messages cannot be obtained. To deal with privacy assessment in these scenarios, we analyze the relationship between the proposed privacy metrics and local centrality properties as an estimation of privacy risk. The results showed that effectiveness centrality can be used as a suitable approximation of the proposed privacy measures., This work was supported in part by the Spanish Government project under Grant TIN2017-89156-R, and in part by the FPI under Grant BES-2015-074498.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Assessing the Effectiveness of a Gamified Social Network for Applying Privacy Concepts: An Empirical Study with Teens

  • Alemany-Bordera, José
  • Del Val Noguera, Elena|||0000-0002-1279-3429
  • García-Fornes, A|||0000-0003-4482-8793
[EN] The concept of privacy in online social networks (OSNs) is a challenge, especially for teenagers. Previous works deal with teaching about privacy using educational online content, and media literacy. However, these tools do not necessarily promote less risky behaviors, and do not allow the assessment of users' behavior after the learning period. Moreover, few research studies about the effects of social gamification have been performed for this population segment (i.e., teenagers). To address this problem in this article, we propose the use of gamification in an OSN called Pesedia to facilitate the teaching/learning process, and assess its effectiveness in promoting suitable privacy behaviors. We tested our proposal comparing teenagers' performance in two editions of a course about social networks, and privacy (with, and without gamification) for one month. We measured the impact of gamification in the participants' behaviors toward privacy concepts as a consequence of the privacy teaching/learning process, and the participants' engagement in the educational process. The results show that there are significant differences in participants' behavior regarding privacy, and engagement in the gamified social network. Moreover, there is also a significant difference in participants' engagement for the gamified male participants. The gamified social network proposed in this article may be relevant, and useful for educators who wish to develop, and enhance teenagers' privacy skills, or for a broader base of aspects related to the development of digital competences, and technology in education., This work was supported in part by the Spanish Government Project TIN2017-89156-R, and in part by the FPI under Grant BFS-2015-074498. (Corresponding author: Elena Del Vol.)
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Tesis doctoral (DoctoralThesis)

MAS-based affective state analysis for user guiding in on-line social environments

  • Aguado Sarrió, Guillem
[ES] Recientemente, hay una fuerte y creciente influencia de aplicaciones en línea en nuestro día a día. Más concretamente las redes sociales se cuentan entre las plataformas en línea más usadas, que permiten a usuarios comunicarse e interactuar desde diferentes partes del mundo todos los días. Dado que estas interacciones conllevan diferentes riesgos, y además los adolescentes tienen características que los hacen más vulnerables a ciertos riesgos, es deseable que el sistema pueda guiar a los usuarios cuando se encuentren interactuando en línea, para intentar mitigar la probabilidad de que caigan en uno de estos riesgos. Esto conduce a una experiencia en línea más segura y satisfactoria para usuarios de este tipo de plataformas. El interés en aplicaciones de inteligencia artificial capaces de realizar análisis de sentimientos ha crecido recientemente. Los usos de la detección automática de sentimiento de usuarios en plataformas en línea son variados y útiles. Se pueden usar polaridades de sentimiento para realizar minería de opiniones en personas o productos, y así descubrir las inclinaciones y opiniones de usuarios acerca de ciertos productos (o ciertas características de ellos), para ayudar en campañas de marketing, y también opiniones acerca de personas como políticos, para descubrir la intención de voto en un periodo electoral, por ejemplo. En esta tesis, se presenta un Sistema Multi-Agente (SMA), el cual integra agentes que realizan diferentes análisis de sentimientos y de estrés usando texto y dinámicas de escritura (usando análisis unimodal y multimodal), y utiliza la respuesta de los analizadores para generar retroalimentación para los usuarios y potencialmente evitar que caigan en riesgos y difundan comentarios en plataformas sociales en línea que pudieran difundir polaridades de sentimiento negativas o niveles altos de estrés. El SMA implementa un análisis en paralelo de diferentes tipos de datos y generación de retroalimentación a través del uso de dos mecanismos diferentes. El primer mecanismo se trata de un agente que realiza generación de retroalimentación y guiado de usuarios basándose en un conjunto de reglas y la salida de los analizadores. El segundo mecanismo es un módulo de Razonamiento Basado en Casos (CBR) que usa no solo la salida de los analizadores en los mensajes del usuario interactuando para predecir si su interacción puede generar una futura repercusión negativa, sino también información de contexto de interacciones de usuarios como son los tópicos sobre los que hablan o información sobre predicciones previas en mensajes escritos por la gente que conforma la audiencia del usuario. Se han llevado a cabo experimentos con datos de una red social privada generada en laboratorio con gente real usando el sistema en tiempo real, y también con datos de Twitter.com para descubrir cuál es la eficacia de los diferentes analizadores implementados y del módulo CBR al detectar estados del usuario que se propagan más en la red social. Esto conlleva descubrir cuál de las técnicas puede prevenir mejor riesgos potenciales que los usuarios pueden sufrir cuando interactúan, y en qué casos. Se han encontrado diferencias estadísticamente significativas y la versión final del SMA incorpora los analizadores que mejores resultados obtuvieron, un agente asesor o guía basado en reglas y un módulo CBR. El trabajo de esta tesis pretende ayudar a futuros desarrolladores de sistemas inteligentes a crear sistemas que puedan detectar el estado de los usuarios interactuando en sitios en línea y prevenir riesgos que los usuarios pudiesen enfrentar. Esto propiciaría una experiencia de usuario más segura y satisfactoria., [CA] Recentment, hi ha una forta i creixent influència d'aplicacions en línia en el nostre dia a dia, i concretament les xarxes socials es compten entre les plataformes en línia més utilitzades, que permeten a usuaris comunicar-se i interactuar des de diferents parts del món cada dia. Donat que aquestes interaccions comporten diferents riscos, i a més els adolescents tenen característiques que els fan més vulnerables a certs riscos, seria desitjable que el sistema poguera guiar als usuaris mentre es troben interactuant en línia, per així poder mitigar la probabilitat de caure en un d'aquests riscos. Açò comporta una experiència en línia més segura i satisfactòria per a usuaris d'aquest tipus de plataformes. L'interés en aplicacions d'intel·ligència artificial capaces de realitzar anàlisi de sentiments ha crescut recentment. Els usos de la detecció automàtica de sentiments en usuaris en plataformes en línia són variats i útils. Es poden utilitzar polaritats de sentiment per a realitzar mineria d'opinions en persones o productes, i així descobrir les inclinacions i opinions d'usuaris sobre certs productes (o certes característiques d'ells), per a ajudar en campanyes de màrqueting, i també opinions sobre persones com polítics, per a descobrir la intenció de vot en un període electoral, per exemple. En aquesta tesi, es presenta un Sistema Multi-Agent (SMA), que integra agents que implementen diferents anàlisis de sentiments i d'estrés utilitzant text i dinàmica d'escriptura (utilitzant anàlisi unimodal i multimodal), i utilitza la resposta dels analitzadors per a generar retroalimentació per als usuaris i potencialment evitar que caiguen en riscos i difonguen comentaris en plataformes socials en línia que pogueren difondre polaritats de sentiment negatives o nivells alts d'estrés. El SMA implementa una anàlisi en paral·lel de diferents tipus de dades i generació de retroalimentació a través de l'ús de dos mecanismes diferents. El primer mecanisme es tracta d'un agent que realitza generació de retroalimentació i guia d'usuaris basant-se en un conjunt de regles i l'eixida dels analitzadors. El segon mecanisme és un mòdul de Raonament Basat en Casos (CBR) que utilitza no solament l'eixida dels analitzadors en els missatges de l'usuari per a predir si la seua interacció pot generar una futura repercussió negativa, sinó també informació de context d'interaccions d'usuaris, com són els tòpics sobre els quals es parla o informació sobre prediccions prèvies en missatges escrits per la gent que forma part de l'audiència de l'usuari. S'han realitzat experiments amb dades d'una xarxa social privada generada al laboratori amb gent real utilitzant el sistema implementat en temps real, i també amb dades de Twitter.com per a descobrir quina és l'eficàcia dels diferents analitzadors implementats i del mòdul CBR en detectar estats de l'usuari que es propaguen més a la xarxa social. Açò comporta descobrir quina de les tècniques millor pot prevenir riscos potencials que els usuaris poden sofrir quan interactuen, i en quins casos. S'han trobat diferències estadísticament significatives i la versió final del SMA incorpora els analitzadors que millors resultats obtingueren, un agent assessor o guia basat en regles i un mòdul CBR. El treball d'aquesta tesi pretén ajudar a futurs dissenyadors de sistemes intel·ligents a crear sistemes que puguen detectar l'estat dels usuaris interactuant en llocs en línia i prevenir riscos que els usuaris poguessen enfrontar. Açò propiciaria una experiència d'usuari més segura i satisfactòria., [EN] In the present days, there is a strong and growing influence of on-line applications in our daily lives, and concretely Social Network Sites (SNSs) are one of the most used on-line social platforms that allow users to communicate and interact from different parts of the world every day. Since this interaction poses several risks, and also teenagers have characteristics that make them more vulnerable to certain risks, it is desirable that the system could be able to guide users when interacting on-line, to try and mitigate the probability of incurring one of those risks. This would in the end lead to a more satisfactory and safe experience for the users of such on-line platforms. Recently, interest in artificial intelligence applications being able to perform sentiment analysis has risen. The uses of detecting the sentiment of users in on-line platforms or sites are variated and rewarding. Sentiment polarities can be used to perform opinion mining on people or products, and discover the inclinations and opinions of users on certain products (or certain features of them) to help marketing campaigns, and also on people such as politics, to discover the voting intention for example in electoral periods. In this thesis, a Multi-Agent System (MAS) is presented, which integrates agents that perform different sentiment and stress analyses using text and keystroke dynamics data (using both unimodal and multi-modal analysis). The MAS uses the output of the analyzers for generating feedback for users and potentially avoids them from incurring risks and spreading comments in on-line social platforms that could lead to the spread of negative sentiment or high-stress levels. Moreover, the MAS incorporates parallelized analyses of different data types and feedback generation via the use of two different mechanisms. On the one hand, a rule-based advisor agent has been implemented, that generates feedback or guiding for users based on the output of the analyzers and a set of rules. On the other hand, a Case-Based Reasoning (CBR) module that uses not only the output of the different analyzers on the messages of the user interacting, but also context information from user interactions such as the topics being talked about or information about the previous states detected on messages written by people in the audience of the user. Experiments with data from a private SNS generated in a laboratory with real people using the system in real-time, and also with data from Twitter.com have been performed to ascertain the efficacy of the different analyzers implemented and the CBR module on detecting states of the user that propagate more in the network, which leads to discovering which of the techniques is able to better prevent potential risks that users could face when interacting, and in which cases. Significant differences were found and the final version of the MAS incorporates the best-performing analyzer agents, a rule-based advisor agent, and a CBR module. In the end, this thesis aims to help intelligent systems developers to build systems that are able to detect the state of users interacting in on-line sites and prevent risks that they could face, leading to a more satisfactory and safe user experience., This thesis was funded by the following research projects: Privacy in Social Educational Environments during Child-hood and Adolescence (PESEDIA), Ministerio de Economia y Empresa (TIN2014-55206-R) and Intelligent Agents for Privacy Advice in Social Networks (AI4PRI), Ministerio de Economia y Empresa (TIN2017-89156-R)
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Metrics for Privacy Assessment When Sharing Information in Online Social Networks

  • Alemany, J.
  • Del Val, E.
  • Alberola, J.M.
  • Garcia-Fornes, A.
Privacy risk in Online Social Networks has become an important social concern. Users, with different perceptions of risk, share information without considering the audience that has access to the information disclosed or how far a publication will go. According to this, we propose two metrics (Audience and Reachability) based on information flows and friendship layers that indicate the privacy risk of sharing information, addressing the posts'' scope and invisible audience. We assess these metrics through agent simulations in well-known models of networks. The findings show a strong relationship between metrics and structural centrality network properties. We also studied scenarios where there is no previous information about users activity or the information about the traces of the messages cannot be obtained. To deal with privacy assessment in these scenarios, we analyze the relationship between the proposed privacy metrics and local centrality properties as an estimation of privacy risk. The results showed that effectiveness centrality can be used as a suitable approximation of the proposed privacy measures.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Assessing the Effectiveness of a Gamified Social Network for Applying Privacy Concepts: An Empirical Study with Teens

  • Alemany, J.
  • Val, E.D.
  • Garcia-Fornes, A.
The concept of privacy in online social networks (OSNs) is a challenge, especially for teenagers. Previous works deal with teaching about privacy using educational online content, and media literacy. However, these tools do not necessarily promote less risky behaviors, and do not allow the assessment of users' behavior after the learning period. Moreover, few research studies about the effects of social gamification have been performed for this population segment (i.e., teenagers). To address this problem in this article, we propose the use of gamification in an OSN called Pesedia to facilitate the teaching/learning process, and assess its effectiveness in promoting suitable privacy behaviors. We tested our proposal comparing teenagers' performance in two editions of a course about social networks, and privacy (with, and without gamification) for one month. We measured the impact of gamification in the participants' behaviors toward privacy concepts as a consequence of the privacy teaching/learning process, and the participants' engagement in the educational process. The results show that there are significant differences in participants' behavior regarding privacy, and engagement in the gamified social network. Moreover, there is also a significant difference in participants' engagement for the gamified male participants. The gamified social network proposed in this article may be relevant, and useful for educators who wish to develop, and enhance teenagers' privacy skills, or for a broader base of aspects related to the development of digital competences, and technology in education.
Zaguán. Repositorio Digital de la Universidad de Zaragoza
Publicación Artículo científico (JournalArticle)

Normative emotional agents: A viewpoint paper

  • Argente, Estefania
  • Val, E. Del
  • Pérez-García, D.
  • Botti, V.
Human social relationships imply conforming to the norms, behaviors, and cultural values of the society, but also socialization of emotions, to learn how to interpret and show them. In multiagent systems, much progress has been made in the analysis and interpretation of both emotions and norms. Nonetheless, the relationship between emotions and norms has hardly been considered and most normative agents do not consider emotions, or vice-versa. In this article, we provide an overview of relevant aspects within the area of normative agents and emotional agents. First we focus on the concept of norm, the different types of norms, its life cycle and a review of multiagent normative systems. Second, we present the most relevant theories of emotions, the life cycle of an agent’s emotions, and how emotions have been included through computational models in multiagent systems. Next, we present an analysis of proposals that integrate emotions and norms in multiagent systems. From this analysis, four relationships are detected between norms and emotions, which we analyze in detail and discuss how these relationships have been tackled in the reviewed proposals. Finally, we present a proposal for an abstract architecture of a Normative Emotional Agent that covers these four norm-emotion relationships