Fabiana Vernero

Mirko Lai, Alessandra Urbinati, Simona Frenda et al.

Current research primarily focuses on model performance, while comparatively less attention has been devoted to uncertainty estimation, particularly in settings where LLMs are increasingly used to generate annotated data. We introduce a framework combining conformal prediction with Collaborative Filtering-style annotators' representation to model LLM behavior in relation to human annotators and to analyze patterns of agreement and disagreement. Using Non-Conformity Scores, we introduce the Ghost Prediction metric and the Ghost Annotator representation to quantify cases in which model predictions diverge from all available human annotations. We compute cosine similarity measures to explore differences in model behavior across sociodemographic axes. We evaluated four LLMs of different size and families across four content moderation datasets. Our finding shows that while we find that all models uncertainty increases with annotator disagreement, larger models tend to be more confident in the classification of texts that are not aligned with any human annotation. Finally, the Ghost Annotator framework reveals a consistent and robust pattern of demographic misalignment, suggesting a structural bias likely rooted in pretraining corpora.

Federica Cena, Cristina Gena, Claudio Mattutino et al.

In this paper, we report on three experiments we have carried out in the context of the EMPATHY project, with the aim of helping users make better configuration choices in a smart home environment, and discuss our results. We found that there are psychological traits, such as Need for Cognition, which influence the way individuals tend to use recommendations, that there are non obvious relationships between the perceived usefulness of recommendations in different domains and individuals' ability to exploit suggestions on configuration choices, and that detailed, easy-to-understand security explanations are more persuasive than simple security warnings, when it comes to make decisions on the applicability of rules which might cause privacy and security risks.

Cristina Gena, Fabiana Vernero, Claudio Mattutino et al.

Interest in and development of touchless gestural interfaces has recently exploded, fueled by the diffusion of both commercial midair gesture platforms and public interactive displays. This paper focuses on an application based on Microsoft Kinect that allows students to browse a university website, hosted on a public display, through simple gestures. We present two empirical evaluations where we evaluated how users react to this new way of interaction. In addition to confirming the current lack of standards, our results provide some inspiration for the design of touchless interaction.

Antonio Lieto, Fabiana Vernero

In the last decade Human-Computer Interaction (HCI) has started to focus attention on forms of persuasive interaction where computer technologies have the goal of changing users behavior and attitudes according to a predefined direction. In this work, we hypothesize a strong connection between logical fallacies (forms of reasoning which are logically invalid but cognitively effective) and some common persuasion strategies adopted within web technologies. With the aim of empirically evaluating our hypothesis, we carried out a pilot study on a sample of 150 e-commerce websites.