Calibration is the degree to which a recommender system is able to match the distribution of a certain item attribute among the items consumed by a user with their respective recommendations.Recent work suggests that many recommenders tend to provide miscalibrated recommendations.Furthermore, most approaches aimed at improving calibration adopt the post-processing paradigm, making them computationally costly at the inference time.This work proposes CaliTune, a fine-tuning approach applied to collaborative filtering based recommenders to allow them generate better calibrated recommendations without relying on costly post-processing.We compare CaliTune to an established post-processing approach on two backbone models and datasets from movie and music domains, focusing on popularity calibration.Our results suggest that CaliTune can offer a competitive accuracy-calibration trade-off in several settings, particularly when the backbone model exhibits high miscalibration and accuracy remains important, making it a promising inference-efficient alternative in such cases.

The widespread adoption of electronic health records has generated a vast amount of patient-related data, mostly presented in the form of unstructured text, which could be used for document retrieval. However, querying these texts in full could present challenges due to their unstructured and lengthy nature, as they may contain noise or irrelevant terms that can interfere with the retrieval process. Recently, large language models (LLMs) have revolutionized natural language processing tasks. However, despite their promising capabilities, their use in the medical domain has raised concerns due to their lack of understanding, hallucinations, and reliance on outdated knowledge. To address these concerns, we evaluate a Retrieval Augmented Generation (RAG) approach that integrates medical knowledge graphs with LLMs to support query refinement in medical document retrieval tasks. Our initial findings from experiments using two benchmark TREC datasets demonstrate that knowledge graphs can effectively ground LLMs in the medical domain.

Hate speech involves harmful expressions that either directly attack or endorse hatred toward a group or individual, often based on identity characteristics such as ethnicity, religion, or sexual orientation. The lack of regulation regarding hate speech raises concerns about the vulnerability of targeted groups and the potential escalation of violence. Detecting and eliminating hate speech from social media is crucial to prevent harm, negative psychological effects, and further attacks. The complexity of identifying hate speech lies in the indirect or implicit nature of its expressions, which can be influenced by context and the use of sarcasm or irony. Detection becomes even more challenging in multimodal scenarios, such as memes, where text and images are combined to convey meaning. To address this problem, this study analyzes the design of a multimodal hate speech detection technique that integrates textual and visual dimensions. This study explores how to define the textual and visual representations of memes and explores fusion strategies. The preliminary experiments conducted over publicly available datasets achieved encouraging results when compared to more complex approaches in the literature, while also highlighting the challenges of the task.

In today's information-saturated social media environment, it is essential to prioritize the verification of potentially false or misleading claims. This need has led to the development of fact-checking, a process dedicated to verifying the truthfulness of statements. Given the limited human resources available to scrutinize all online claims, it is crucial to identify the most critical ones to verify. Therefore, a (semi-)automated system capable of detecting the most urgent and relevant claims for verification is needed. To address this challenge, we evaluate an approach based on Natural Language Processing and Machine Learning techniques. We explore lexical features, embedding models, LLMs, and traditional classification techniques to develop an automated system to classify statements according to their relevance for verification (i.e., their check-worthiness). Our evaluation is based on data collections including checkable statements extracted from tweets and political speeches. Embeddingand LLM-based techniques showed great potential to improve the performance of the verification process by effectively prioritizing the most critical and relevant statements for verification.

Gender bias is a common and often neglected issue in legal documents. It arises from perceptions or prejudices about the characteristics of a group, or the roles individuals should play in society. This bias can significantly impact the reasoning or outcomes of legal processes, such as judicial rulings. To ensure equal treatment for all individuals, it is crucial to effectively reduce this bias. The first step to reduce bias is to define approaches that can identify manifestations of gender bias. However, these manifestations are not usually easily detectable in text (e.g., through keywords) as they often require detailed contextual analysis, typically done manually by experts. This paper addresses this issue by leveraging natural language processing and machine learning techniques to automate parts of the analysis. Specifically, it proposes a processing pipeline based on text embeddings, binary classification, and the use of large language models (LLMs) to explain classification results. An initial evaluation on a set of judicial rulings shows promising results in terms of precision and recall, along with qualitative insights into the potential of these techniques in the legal domain.

Recommender systems play a crucial role in how users consume information, with group recommendation receiving considerable attention. Ensuring fairness in group recommender systems entails providing recommendations that are useful and relevant to all group members rather than solely reflecting the majority’s preferences, while also addressing fairness concerns related to sensitive attributes (e.g., gender). Recently, the advancements on Large Language Models (LLMs) have enabled the development of new kinds of recommender systems. However, LLMs can perpetuate social biases present in training data, posing risks of unfair outcomes and harmful impacts. We investigated LLMs impact on group recommendation fairness, establishing and instantiating a framework that encompasses group definition, sensitive attribute combinations, and evaluation methodology. Our findings revealed the interaction patterns between sensitive attributes and LLMs and how they affected recommendation. This study advances the understanding of fairness considerations in group recommendation systems, laying the groundwork for future research.

Group recommenders aim to provide recommendations that satisfy the collective preferences of multiple users, a challenging task due to the diverse individual tastes and conflicting interests to be balanced. This is often accomplished by using aggregation techniques that select items on which the group can agree. Traditional aggregators struggle with these complexities, as items are chosen independently, leading to sub-optimal recommendations lacking diversity, novelty, or fairness. In this paper, we propose an aggregation technique that leverages Monte Carlo Tree Search (MCTS) to enhance group recommendations. MCTS is used to explore and evaluate candidate recommendation sequences to optimize overall group satisfaction. We also investigate the integration of MCTS with LLMs aiming at better understanding interactions between user preferences and recommendation sequences to inform the search. Experimental evaluations, although preliminary, showed that our proposal outperforms existing aggregation techniques in terms of relevance and beyond-accuracy aspects of recommendations. The LLM integration achieved positive results for recommendations’ relevance. Overall, this work highlights the potential of heuristic search techniques to tackle the complexities of group recommendations.

Recommender systems play a crucial role in how users consume information and establish new social relations. However, different factors (such as the data collection process, the designed recommendation model, or even the interpretation of findings) could make recommenders (unintendedly) prone to biases, favouring certain user groups or items, thus resulting in unfair outcomes. Recommenders also face fairness criticism for inducing filter bubbles, echo chambers, and, more generally, facilitating opinion manipulation. In this work, we study the impact of user recommender systems on fairness. To this end, we carry out a user recommendation task on a politically polarized Twitter data collection. Then, we evaluate how the different politically aligned user groups experience recommendation quality. Finally, we explore causal models to identify data and model-related features that could affect the fairness of recommender outcomes. Our study shows that political alignment is associated with the unfairness of recommenders affecting not only the relevance of recommendations, but also their diversity and the resulting interaction patterns.

When developing Java Script (JS) applications, the assessment and selection of JS packages have become challenging for developers due to the growing number of technology options available. Given a technology need, a common developers' strat-egy is to query Web repositories via search engines (e.g., NPM, Google) and shortlist candidate JS packages. However, these engines might return a long list of results. Furthermore, these results should be ranked according to the developer's criteria. To address these problems, we developed a recommender system called AIDT that assists developers in the package selection task. AIDT relies on meta-search and machine learning techniques to infer the relevant packages for a query. An initial evaluation of AIDT showed good search effectiveness. Recently, the emergence of ChatGPT has opened new opportunities for this kind of assistants, as reported by some experiments. Anyway, human developers should judge whether the recommendations (e.g., JS packages) of these tools are fit to purpose. In this paper, we report on a user study in which we used both AIDT and ChatGPT on a sample of JS-related queries, compared their results, and also validated them against developers' criteria and expectations for the task. Our initial findings show that ChatGPT is not yet on par with AIDT or even human efforts for the task at hand, but the model is flexible to be improved and furthermore, it can provide good arguments for its package choices.

With the clinical environment becoming more data-reliant, healthcare professionals now have unparalleled access to comprehensive clinical information from numerous sources. Then, one of the main issues is how to avoid overloading practitioners with large amounts of (irrelevant) information while guiding them to the relevant documents for specific patient cases. Additional challenges appear due to the shortness of queries and the presence of long (and maybe noisy) contextual information. This demo presents Re2Dan, a web Retrieval and recommender of Danish medical documents. Re2Dan leverages several techniques to improve the quality of retrieved documents. First, it combines lexical and semantic searches to understand the meaning and context of user queries, allowing the retrieval of documents that are conceptually similar to the user’s query. Second, it recommends similar queries, allowing users to discover related documents and insights. Third, when given contextual information (e.g., from patients’ clinical notes), it suggests medical concepts to expand the user query, enabling a more focused search scope and thus obtaining more accurate recommendations. Preliminary analyses showed the effectiveness of the recommender in improving the relevance and comprehensiveness of recommendations, thereby assisting healthcare professionals in finding relevant information for informed decision-making.

Recommender systems are central to online information consumption and user-decision processes, as they help users find relevant information and establish new social relationships. However, recommenders could also (unintendedly) help propagate misinformation and increase the social influence of the spreading it. In this context, we study the impact of friend recommender systems on the social influence of misinformation spreaders on Twitter. To this end, we applied several user recommenders to a COVID-19 misinformation data collection. Then, we explore what-if scenarios to simulate changes in user misinformation spreading behaviour as an effect of the interactions in the recommended network. Our study shows that recommenders can indeed affect how misinformation spreaders interact with other users and influence them.

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.

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 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.