Search

The topic of algorithmic fairness is of increasing importance to the Human–Computer Interaction research community following accumulating concerns regarding the use and deployment of Artificial Intelligence-based systems. How we conduct research on algorithmic fairness directly influences our inferences and conclusions regarding algorithmic fairness. To better understand the methodological decisions of studies focused on people's perceptions of algorithmic fairness, we systematic analysed relevant papers from the CHI and FAccT conferences. We identified 200 relevant papers published between 1993 and 2022 and assessed their study design, participant sample, and geographical location of participants and authors. Our results highlight that studies are predominantly cross-sectional, cover a wide range of participant roles, and that both authors and participants are primarily from the United States. Based on these findings, we reflect on the potential pitfalls and shortcomings in how the community studies algorithmic fairness. • Mapping of CHI and FAccT papers on algorithmic fairness published 1993–2022. • Over one-third of papers miss details on participant locality and compensation. • Algorithmic fairness research largely restricted to US and other Western countries. • Limited cross-country collaboration, most study samples from a single country. • Remote studies are most common, longitudinal studies are relatively rare. [ABSTRACT FROM AUTHOR]

Abstract Reputation is crucial to enabling human or software agents to select among alternative providers. Although several effective reputation assessment methods exist, they typically distil reputation into a numerical representation, with no accompanying explanation of the rationale behind the assessment. Such explanations would allow users or clients to make a richer assessment of providers, and tailor selection according to their preferences and current context. In this paper, we propose an approach to explain the rationale behind assessments from quantitative reputation models, by generating arguments that are combined to form explanations. Our approach adapts, extends and combines existing approaches for explaining decisions made using multi-attribute decision models in the context of reputation. We present example argument templates, and describe how to select their parameters using explanation algorithms. Our proposal was evaluated by means of a user study, which followed an existing protocol. Our results give evidence that although explanations present a subset of the information of trust scores, they are sufficient to equally evaluate providers recommended based on their trust score. Moreover, when explanation arguments reveal implicit model information, they are less persuasive than scores. [ABSTRACT FROM AUTHOR]

Abstract: While nowadays the most usual way to show emotions in digital contexts is via virtual characters, its use may raise false expectations (the user attributes human abilities to the virtual character). This paper proposes and explores an approach to express emotions which intends to minimize the user's expectations by using a non-anthropomorphic model. Emotions are represented in terms of arousal and valence dimensions. They are visualized in a simple way through the behaviour and appearance of a series of cartoonish clouds. In particular, the arousal value is expressed through the movement of these clouds (controlled by a flocking algorithm), while the valence value is expressed through their degree of darkness. Furthermore, the paper describes a user experiment which investigated whether the arousal and valence expressed by our model are appropriately interpreted by the users or not. The results suggest that movement and darkness are interpreted as arousal and valence respectively and that they are independent of each other. [Copyright &y& Elsevier]

Abstract: Users who visit a web page for the first time select links on the basis of the link anchors and the descriptions of the links that are provided on the page. If these descriptions give an inaccurate impression of the underlying pages, users make incorrect choices. In this paper we present a novel algorithm to automatically detect links with problematic descriptions in hierarchical menus on the basis of usage information stored in the sites’ log files. The algorithm searches the log files for navigation patterns that show that many users make navigation mistakes at certain points in the site. Repeated mistakes indicate that users do not understand the intentions of a link description. The algorithm distinguishes several types of problematic descriptions and provides recommendations for how problematic descriptions of each type can be improved. The findings of the algorithm can help a webmaster of a site to gain insights in the behavior of the sites’ users and to improve the site''s menu accordingly. Evaluation in three domains shows that webmasters perceive the analyses made by the algorithm as very helpful for improving sites. Moreover, a user experiment demonstrates that the adaptations made by the webmasters on the basis of the analyses, significantly reduce the number of navigation mistakes. [Copyright &y& Elsevier]

This paper presents a model, called Scene-Driver, for the re-use of film and television material. We begin by exploring general issues surrounding the ways in which content can be sub-divided into meaningful units for re-use and how criteria might then be applied to the selection and ordering of these units. We also identify and discuss the different means by which a user might interact with the content to create novel and engaging experiences. The Scene-Driver model has been instantiated using content from an animated children's television series called Tiny Planets, which is aimed at children of 5–7-year old. This type of material, being story-based itself, lends itself particularly well to the application of narrative constraints to scene reordering, to provide coherence to the experience of interacting with the content. We propose an interactive narrative-driven game architecture, in which a user generates novel narratives from existing content by placing “domino” like tiles. These tiles act as “glue” between scenes and each tile choice dictates certain properties of the next scene to be shown within a game. There are three different game-types, based on three different ways in which tiles can be matched to scenes. We introduce algorithms for generating legal tile-sets for each of these three game-types, which can be extended to include narrative constraints. This ensures that all novel orderings adhere to a minimum narrative plan, which has been identified based on analysis of the Tiny Planets series and on narrative theories. We also suggest ways in which basic narratives can be enhanced by the inclusion of directorial techniques and by the use of more complex plot structures. In our evaluation studies with children in the target age-range, our game compared favourably with other games that the children enjoyed playing. [Copyright &y& Elsevier]

Abstract: Speed-dependent automatic zooming (SDAZ) has been proposed for standard desktop displays as a means of overcoming problems associated with the navigation of large information spaces. SDAZ combines zooming and panning facilities into a single operation, with the magnitude of both factors dependent on simple user interaction. Previous research indicated dramatic user performance improvements when using the technique for document and map navigation tasks. In this paper, we propose algorithmic extensions to the technique for application on small-screen devices and present a comparative experimental evaluation of user performance with the system and a normative scroll–zoom–pan interface. Users responded positively to the system, particularly in relation to reduced physical navigational workload. However, the reduced screen space reduced the impact of SDAZ in comparison to that reported in previous studies. In fact, for one-dimensional navigation (vertical document navigation) the normative interface out-performed SDAZ. For navigation in two dimensions (map browsing) SDAZ supports more accurate target location, and also produces longer task completion times. Some SDAZ users became lost within the information space and were unable to recover navigational context. We discuss the reasons for these observations and suggest ways in which limitations of SDAZ in the small-screen context may be overcome. [Copyright &y& Elsevier]