A5074802146- Recommender Systems and Techniques
- Decision-Making and Behavioral Economics
- Advanced Bandit Algorithms Research
- Behavioral Health and Interventions
- Environmental Education and Sustainability
- Consumer Market Behavior and Pricing
- Economic and Environmental Valuation
- Digital Marketing and Social Media
- Innovative Human-Technology Interaction
- Explainable Artificial Intelligence (XAI)
- Experimental Behavioral Economics Studies
- Environmental Sustainability in Business
- Music Technology and Sound Studies
- Neuroscience and Music Perception
- Music and Audio Processing
- Ethics and Social Impacts of AI
- Data Visualization and Analytics
- Advanced Text Analysis Techniques
- Intelligent Tutoring Systems and Adaptive Learning
- Mobile Crowdsensing and Crowdsourcing
- Topic Modeling
- Technology Adoption and User Behaviour
- Human-Automation Interaction and Safety
- Image and Video Quality Assessment
- Semantic Web and Ontologies
Eindhoven University of Technology
2016-2025
Arq Psychotrauma Expert Group
2021-2022
Human Computer Interaction (Switzerland)
2006-2020
University of Bergen
2019
Delft University of Technology
2018
Research on recommender systems typically focuses the accuracy of prediction algorithms. Because only partially constitutes user experience a system, this paper proposes framework that takes user-centric approach to system evaluation. The links objective aspects behavior through series perceptual and evaluative constructs (called subjective experience, respectively). Furthermore, it incorporates influence personal situational characteristics experience. This reviews how current literature...
Even though people are attracted by large, high quality recommendation sets, psychological research on choice overload shows that choosing an item from sets containing many attractive items can be a very difficult task. A web-based user experiment using matrix factorization algorithm applied to the MovieLens dataset was used investigate effect of set size (5 or 20 items) and (low high) perceived variety, attractiveness, difficulty satisfaction with chosen item. The results show larger only...
Recent developments in user evaluation of recommender systems have brought forth powerful new tools for understanding what makes recommendations effective and useful. We apply these methods to understand how users evaluate recommendation lists the purpose selecting an algorithm finding movies. This paper reports on experiment which we asked compare produced by three common collaborative filtering algorithms dimensions novelty, diversity, accuracy, satisfaction, degree personalization, select...
Decision research has experienced a shift from simple algebraic theories of choice to an appreciation mental processes underlying choice. A variety process-tracing methods helped researchers test these process explanations. Here, we provide survey methods, including specific examples for subject reports, movement-based measures, peripheral psychophysiology, and neural techniques. We show how can inform phenomena as varied attention, emotion, strategy use, understanding correlates. Two...
This paper compares five different ways of interacting with an attribute-based recommender system and shows that types users prefer interaction methods. In online experiment energy-saving the methods are compared in terms perceived control, understandability, trust system, user interface satisfaction, effectiveness choice satisfaction. The comparison takes into account several characteristics, namely domain knowledge, trusting propensity persistence. results show most (and particularly...
This report documents the program and outcomes of Dagstuhl Seminar 23031 "Frontiers Information Access Experimentation for Research Education", which brought together 38 participants from 12 countries. The seminar addressed technology-enhanced information access (information retrieval, recommender systems, natural language processing) specifically focused on developing more responsible experimental practices leading to valid results, both research as well scientific education. featured a...
The adoption of recommender systems (RSs) in various domains has become increasingly popular, but concerns have been raised about their lack transparency and interpretability. While significant advancements made creating explainable RSs, there is still a shortage automated approaches that can deliver meaningful contextual human-centered explanations. Numerous researchers evaluated explanations based on human-generated recommendations to address this gap. However, such do not scale for...
Behaviorism is the currently-dominant paradigm for building and evaluating recommender systems. Both operation evaluation of system applications are most often driven by analyzing behavior users. In this paper, we argue that listening to what users say about items recommendations they like, control wish exert on output, ways in which perceive not just observing do will enable important developments future We provide both philosophical pragmatic motivations idea, describe various points...
As recommender systems are increasingly deployed in the real world, they not merely tested offline for precision and coverage, but also "online" with test users to ensure good user experience. The evaluation of recommenders is however complex resource-consuming. We introduce a pragmatic procedure evaluate experience products users, within industry constraints on time budget. Researchers practitioners can employ our approach gain comprehensive understanding their systems.
People like variety and often prefer to choose from large item sets. However, sets can cause a phenomenon called "choice overload": they are more difficult from, as result decision makers less satisfied with their choices. It has been argued that choice overload occurs because contain similar items. To overcome this effect, the present paper proposes increasing diversity of might make them attractive satisfactory, without making much from. purpose, by using structural equation model...
We investigate the usability of humanlike agent-based interfaces for interactive advice-giving systems. In an experiment with a travel advisory system, we manipulate “humanlikeness” agent interface. demonstrate that users more agents try to exploit capabilities were not signaled by system. This severely reduces systems look human but lack humanlikehumanlike (overestimation effect). explain this effect showing form anthropomorphic beliefs (a user's “mental model”) about system: They act...
Recipe websites are becoming increasingly popular to support people in their home cooking. However, most of these prioritize recipes, which tend be unhealthy. Drawing upon research on visual biases and nudges, this paper investigates whether healthy food choices can supported search by depicting attractive images alongside as well re-ranking results health. After modelling the attractiveness recipe images, we asked 239 users for specific online recipes select those they liked most. Our...
Humans increasingly interact with AI systems, and successful interactions rely on individuals trusting such systems (when appropriate). Considering that trust is fragile often cannot be restored quickly, we focus how develops over time in a human-AI-interaction scenario. In 2x2 between-subject experiment, test model accuracy (high vs. low) type of explanation (human-like not) affect time. We study complex decision-making task which estimate jail for 20 criminal law cases advice. Results show...
The primary goal of Recommender Systems is to suggest the most suitable items a user, aligning them with user's interests and needs. RSs are essential for modern e-commerce, helping users discover content products by predicting based on their past behavior. However, success isn't just about advanced algorithms. design user interface good integration human decision-making process equally crucial. A well-designed enhances experience makes recommendations more effective, while poor can lead...
Loss aversion and reference dependence are 2 keystones of behavioral theories choice, but little is known about their underlying cognitive processes. We suggest an additional account for loss that supplements the current value encoding attributes as gains or losses relative to a point, introducing construction account. Value suggests results from biased evaluations during information search comparison develop hypotheses identify influence both accounts examine process-tracing data evidence....
One of the challenges for recommender systems is that users struggle to accurately map their internal preferences external measures quality such as ratings. We study two methods supporting mapping process: (i) reminding user characteristics items by providing personalized tags and (ii) relating rating decisions prior using exemplars. In our study, we introduce interfaces provide these support. also present a set methodologies evaluate efficacy new via experiment. Our results suggest...
People often struggle to find appropriate energy-saving measures take in the household. Although recommender studies show that tailoring a system's interaction method domain knowledge of user can increase energy savings, they did not actually tailor conservation advice itself. We present two large which we support users make an energy-efficient behavioral change by presenting tailored advice. Both systems use one-dimensional, ordinal Rasch scale, orders 79 on their difficulty and link this...
We studied an alternative choice-based interface for preference elicitation during the cold start phase and compared it directly with a standard rating-based interface. In this users started from diverse set covering all movies iteratively narrowed down through matrix factorization latent feature space to smaller sets of items based on their choices. The results show that interface, requires less effort in more satisfying recommendations, showing might be promising candidate alleviating...
Model-agnostic explainable AI tools explain their predictions by means of 'local' feature contributions. We empirically investigate two potential improvements over current approaches. The first one is to always present contributions in terms the contribution outcome that perceived as positive user ("positive framing"). second add "semantic labeling", explains directionality each ("this leads +5% eligibility"), reducing additional cognitive processing steps. In a study, participants evaluated...
Rapid innovations in electronic healthcare and behavior tracking systems are challenging health coaches (dietitians, personal trainers, etc.) to rethink their traditional roles practices. At the same time, many current e-coaching have been developed without explicitly incorporating professionals' perspective into design process. In paper, we present three consecutive qualitative studies, starting from coach's on successful coaching, progressively zooming potential role impact of technology...