Pharmacovigilance performed from social media data is an active research field that contributes to the automatic detection of adverse drug reactions (ADRs) of medications and vaccines. Natural language processing techniques combined with machine learning models are used to perform the challenging task of analyzing heterogeneous short text content. This study explores the application of state-of-the-art transfer learning approaches for classifying Spanish tweets to identify mentions of ADRs as a result of COVID-19 vaccination. We created a corpus of 1332 tweets about COVID-19 post-vaccination adverse reactions and employed language models for text classification. Preliminary results suggest that these models achieve superior performances in terms of F1 score compared to traditional machine learning models.

The COVID-19 pandemic has affected all aspects of society, bringing health hazards and posing challenges to public order, governments, and mental health. This study examines the stages of crisis response and recovery as a sociological problem by operationalising a well-known model of crisis stages in terms of a psycho-linguistic analysis. Based on an extensive collection of Twitter data spanning from March to August 2020 in Argentina, we present a thematic study on the differences in language used in social media posts and look at indicators that reveal the distinctive stages of a crisis and the country response thereof. The analysis was combined with a study of the temporal prevalence of mental health related conversations and emotions. This approach can provide insights for public health policy design to monitor and eventually intervene during the different stages of a crisis, thus improving the adverse mental health effects on the population.

Recommender systems play a central role in online information consumption and user decision-making by leveraging user-generated information at scale to assist users in finding relevant information and establishing new social relationships. Just as recommendation techniques have become powerful tools that are inserted in most social platforms, they could also involuntarily spread unwanted content and other types of online harms. The same fundamental concepts on which these techniques rely make them facilitators of such unwanted diffusion. To increase the user-perceived quality of recommender systems and mitigating the negative effects of the multiple forms of online harms, it is essential to provide recommender systems with harm-aware mechanisms. To further research in this direction, this Second edition of the Workshop on Online Misinformation- and Harm-Aware Recommender Systems (OHARS 2021) aimed at fostering research in recommender systems that can mitigate the negative effects of online harms by fostering the recommendation of safe content and trustworthy users, with a special interest in research tackling the negative effects of the propagation of harmful content referring to the COVID-19 crisis.

Recommender systems serve as mediators of information consumption and propagation. In this role, these systems have been recently criticized for introducing biases and promoting the creation of echo chambers and filter bubbles, thus lowering the diversity of both content and potential new social relations users are exposed to. Some of these issues are a consequence of the fundamental concepts on which recommender systems are based on. Assumptions like the homophily principle might lead users to content that they already like or friends they already know, which can be naïve in the era of ideological uniformity and fake news. A significant challenge in this context is how to effectively learn the dynamic representations of users based on the content they share and their echo chamber or community interactions to recommend potentially relevant and diverse friends from outside the network of influence of the users’ echo chamber. To address this, we devise FRediECH (a Friend RecommenDer for breakIng Echo CHambers), an echo chamber-aware friend recommendation approach that learns users and echo chamber representations from the shared content and past users’ and communities’ interactions. Comprehensive evaluations over Twitter data showed that our approach achieved better performance (in terms of relevance and novelty) than state-of-the-art alternatives, validating its effectiveness.

Recommender systems play an important role in the dissemination and propagation of information. This is particularly true for large scale platforms such as social media, where recommender systems assist users in facilitating access to massive user-generated content by finding relevant information and establishing new social relationships. Just as recommendation techniques are designed to become powerful tools, they could in turn spread online harm. Some of these issues stem from the core concepts and assumptions of recommender systems. Harnessing recommender systems with misinformation- and harm-awareness mechanisms becomes essential not only to mitigate the negative effects of the propagation of harmful content, but also to increase the quality and diversity of recommender systems. To further research in this direction, the Workshop on Online Misinformation- and Harm-Aware Recommender Systems (OHARS 2020) aimed at fostering research in recommender systems that can circumvent the negative effects of online harms by promoting the recommendation of safe content and users.

Software systems naturally evolve, and this evolution often brings design problems that cause system degradation. Architectural smells are typical symptoms of such problems, and several of these smells are related to undesired dependencies among modules. The early detection of these smells is important for developers, because they can plan ahead for maintenance or refactoring efforts, thus preventing system degradation. Existing tools for identifying architectural smells can detect the smells once they exist in the source code. This means that their undesired dependencies are already created. In this work, we explore a forward-looking approach that is able to infer groups of likely module dependencies that can anticipate architectural smells in a future system version. Our approach considers the current module structure as a network, along with information from previous versions, and applies link prediction techniques (from the field of social network analysis). In particular, we focus on dependency-related smells, such as Cyclic Dependency and Hub-like Dependency, which fit well with the link prediction model. An initial evaluation with two open-source projects shows that, under certain considerations, the predictions of our approach are satisfactory. Furthermore, the approach can be extended to other types of dependency-based smells or metrics.

Since their beginnings, social networks have affected the way people communicate and interact with each other. Nowadays, user interactions range from social relations to posting and reading activities, leading to the existence of multiple and complementary information sources or dimensions for characterising user behaviour. The task of community detection could benefit from integrating those multiple sources. However, most techniques disregard the effect of information aggregation, and continue to focus only on one aspect: network topology. This paper aims at providing some insights on how to integrate the multiple and heterogeneous social media information sources characterising user activities and behaviour to optimise the quality of found communities. To that end, diverse consensus strategies to extend techniques designed for a unique information source to multi-dimensional networks are presented and analysed. Experimental evaluation confirmed the benefits of using consensus strategies for leveraging on multiple data dimensions in terms of community quality.

The Service Oriented Computing (SOC) paradigm promotes building new applications by discovering and then invoking services, i.e., software components accessible through the Internet. Discovering services means inspecting registries where textual descriptions of services functional capabilities are stored. To automate this, existing approaches index descriptions and associate users' queries to relevant services. However, the massive adoption of Web-exposed API development practices, specially in large service ecosystems such as the IoT, is leading to evergrowing registries which challenge the accuracy and speed of such approaches. The recent notion of Social Web Services (SWS), where registries not only store service information but also sociallike relationships between users and services opens the door to new discovery schemes. We investigate an approach to discover SWSs that operates on graphs with user-service relationships and employs lightweight topological metrics to assess service similarity. Then, "socially" similar services, which are determined exploiting explicit relationships and mining implicit relationships in the graph, are clustered via exemplar-based clustering to ultimately aid discovery. Experiments performed with the ProgrammableWeb.com registry, which is at present the largest SWS repository with over 15k services and 140k user-service relationships, show that pure topology-based clustering may represent a promising complement to content-based approaches, which in fact are more time-consuming due to text processing operations.

Large scale multi-label learning, i.e. the problem of determining the associated set of labels for an instance, is gaining relevance in recent years due to the emergence of several real-world applications. Most notably, the exponential growth of the Social Web where a resource can be labeled by mil lions of users using one or more tags, i.e. a resource can be associated to several labels at the same time. A well-known approach for multi-label classification is the Binary Relevance (BR) algorithm which trains a binary classifier for each label independently. However, the serial implementation of BR is not suitable for medium or large datasets due to the time and computational resources required for training. For example, training classifiers for mid-size datasets using MULAN implementation of BR might take several weeks. This paper discusses a parallel implementation of the MULAN BR technique that harnesses the computational power of nowadays multi-core processors. Our implementation presents a speed-up in the training phase of up to 12 times when compared to the original MULAN implementation. In addition, the cross-validation technique of MULAN had huge RAM requirements, making it unusable with large datasets. Therefore, we have overcome this limitation by using compact data structures and taking advantage of disk caching. We have also compared our implementation against scikit-learn, a popular tool for data mining and data analysis, showing significant improvements in speed-up.

The accurate suggestion of interesting friends arises as a crucial issue in recommendation systems. The selection of friends or followees responds to several reasons whose importance might differ according to the characteristics and preferences of each user. Furthermore, those preferences might also change over time. Consequently, understanding how friends or followees are selected emerges as a key design factor of strategies for personalized recommendations. In this work, we argue that the criteria for recommending followees needs to be adapted and combined according to each user's behavior, preferences, and characteristics. A method is proposed for adapting such criteria to the characteristics of the previously selected followees. Moreover, the criteria can evolve over time to adapt to changes in user behavior, and broaden the diversity of the recommendation of potential followees based on novelty. Experimental evaluation showed that the proposed method improved precision results regarding static criteria weighting strategies and traditional rank aggregation techniques.

Collaborative tagging systems, also known as folksonomies, have grown in popularity over the Web on account of their simplicity to organize several types of content (e.g., Web pages, pictures, and video) using open-ended tags. The rapid adoption of these systems has led to an increasing amount of users providing information about themselves and, at the same time, a growing and rich corpus of social knowledge that can be exploited by recommendation technologies. In this context, tripartite relationships between users, resources, and tags contained in folksonomies set new challenges for knowledge discovery approaches to be applied for the purposes of assisting users through recommendation systems. This review aims at providing a comprehensive overview of the literature in the field of folksonomy-based recommender systems. Current recommendation approaches stemming from fields such as user modeling, collaborative filtering, content, and link-analysis are reviewed and discussed to provide a starting point for researchers in the field as well as explore future research lines.

SUMMARY Social tagging systems allow users to easily create, organize, and share collections of Web resources in a collaborative fashion. Videos, pictures, research papers, and Web pages are shared and annotated in sites such as Del.icio.us, CiteULike, or Flickr, among others. The rising popularity of these systems leads to a constant increase in the number of users actively publishing and annotating resources and, consequently, an exponential growth in the amount of data contained in their folksonomies, the underlying data structure of tagging systems. In turn, the user task of discovering interesting resources becomes more and more difficult and time‐consuming. In this paper, the problem of filtering resources from social tagging systems according to individual user interests using purely tagging data is studied. One‐class support vector machine classification is evaluated as a means to identify relevant information for users based exclusively on positive examples of their information preferences. It is assumed that users express their interest on resources belonging to a folksonomy by assigning tags to them, whereas there is no straightforward method to collect uninterestingness judgments. Filtering interesting resources based on social tags is an important benefit of exploiting the collective knowledge generated by tagging activities of Web communities. In this paper, the results achieved with tag‐based classification are compared with those obtained using more traditional information sources such as the full text of Web pages. Experimental evaluation showed that tag‐based classifiers outperformed those learned using the text of documents as well as other content‐related sources. Moreover, tag‐based classification becomes essential for folksonomies in which no additional content is available because of the nature of resources being stored (e.g., tagging of photos or videos). Copyright © 2012 John Wiley & Sons, Ltd.

For personal information agents, user profiles have to represent user interests and preferences in order to satisfy long-term information needs. An implicit assumption in user-profiling is the existence of persistent interests which, however, might suffer some changes over time. Each time the interests of a user change, his profile becomes inaccurate and the predictive quality decreases. Adaptation of user profiles is, therefore, an essential requirement for personal agents that need to be capable of adjusting their behavior quickly in order to shorten the period of reduced predictive quality. In this paper, a user-profiling technique named WebProfiler, which learns a hierarchical representation of user interests using conceptual clustering, is augmented with an adaptation strategy based on relevance feedback and time-based forgetting in order to deal with drifting interests. We empirically evaluate the performance of this strategy by analyzing its behavior on multiple scenarios of interest drifts and shifts.

The motivation behind personal information agents resides in the enormous amount of information available on the Web, which has created a pressing need for effective personalized techniques. In order to assists Web search these agents rely on user profiles modeling information preferences, interests and habits that help to contextualize user queries. In communities of people with similar interests, collaboration among agents fosters knowledge sharing and, consequently, potentially improves the results of individual agents by taking advantage of the knowledge acquired by other agents. In this paper, we propose an agent-based recommender system for supporting collaborative Web search in groups of users with partial similarity of interests. Empirical evaluation showed that the interaction among personal agents increases the performance of the overall recommender system, demonstrating the potential of the approach to reduce the burden of finding information on the Web.