Sensors · 2022 · 37 citations · 28 references
ComputingComputer ScienceArtificial IntelligenceHuman-computer InteractionUser Experience DesignHuman-machine InterfacesInterface DesignUsability EngineeringUsability EvaluationUsability TestingEvaluationEvaluation MetricsExperimental Data CollectionHuman-computer Interface DesignIntelligent User InterfacesInterface EvaluationInterface UsabilityQuestionnaire-based Usability Assessment
Intelligent user interfaces aim to enhance human–computer interaction by improving user experience and usability through artificial intelligence. The study seeks to locate, evaluate, and synthesize state‑of‑the‑art research on intelligent user interfaces, emphasizing their evaluation. The authors conducted a systematic mapping of 211 studies published from 2012 to 2022. The mapping shows that IUI research is predominantly situated in HCI and software engineering, defines IUI through adaptation, representation, and intelligence, evaluates it mainly with experiments and questionnaires, reports usability or UX improvements in 81 % of studies, yet identifies a shortage of dedicated evaluation tools, methods, and metrics.
Intelligent user interfaces (IUI) are driven by the goal of improvement in human–computer interaction (HCI), mainly improving user interfaces’ user experience (UX) or usability with the help of artificial intelligence. The main goal of this study is to find, assess, and synthesize existing state-of-the-art work in the field of IUI with an additional focus on the evaluation of IUI. This study analyzed 211 studies published in the field between 2012 and 2022. Studies are most frequently tied to HCI and SE domains. Definitions of IUI were observed, showing that adaptation, representation, and intelligence are key characteristics associated with IUIs, whereas adaptation, reasoning, and representation are the most commonly used verbs in their description. Evaluation of IUI is mainly conducted with experiments and questionnaires, though usability and UX are not considered together in evaluations. Most evaluations (81% of studies) reported partial or complete improvement in usability or UX. A shortage of evaluation tools, methods, and metrics, tailored for IUI, is noticed. Most often, empirical data collection methods and data sources in IUI evaluation studies are experiment, prototype development, and questionnaire.
28
Systematic Mapping Studies in Software Engineering
Kai Petersen, Robert Feldt, Shahid Mujtaba et al. · Electronic workshops in computing · 2008 · 3.1K citations · Full text
Trends and Trajectories for Explainable, Accountable and Intelligible Systems
Ashraf Abdul, Jo Vermeulen, Danding Wang et al. · 2018 · 723 citations