In the absence of interaction history, image recommendations often depend on content-based approaches. Prompted by user queries in natural language, such systems rank items based on the similarity between textual and visual features. However, these approaches typically rely on static queries and do not offer alternative feedback mechanisms. In this paper, we present VisualReF: an interactive image retrieval prototype that introduces visual relevance feedback through fine-grained user annotations. Built on vision-language models (VLMs) for retrieval, our system allows users to label relevant and irrelevant regions in retrieved images. These regions are captioned using a generative vision-language model to refine the query vector. Our work bridges the gap between conventional static image retrieval and interactive, user-guided search by introducing visual relevance feedback. Finally, our prototype contributes to the field of visual recommendation by empowering researchers with practical tools for: (i) collecting region-level visual relevance signals from users, (ii) supporting integration of human feedback into interactive search pipelines, and (iii) explaining how the relevance feedback model perceives user input.

Large Language Models (LLMs) are increasingly being implemented as joint decision-makers and explanation generators for Group Recommender Systems (GRS). In this paper, we evaluate these recommendations and explanations by comparing them to social choice-based aggregation strategies. Our results indicate that LLM-generated recommendations often resembled those produced by Additive Utilitarian (ADD) aggregation. However, the explanations typically referred to averaging ratings (resembling but not identical to ADD aggregation). Group structure, uniform or divergent, did not impact the recommendations. Furthermore, LLMs regularly claimed additional criteria such as user or item similarity, diversity, or used undefined popularity metrics or thresholds. Our findings have important implications for LLMs in the GRS pipeline as well as standard aggregation strategies. Additional criteria in explanations were dependent on the number of ratings in the group scenario, indicating potential inefficiency of standard aggregation methods at larger item set sizes. Additionally, inconsistent and ambiguous explanations undermine transparency and explainability, which are key motivations behind the use of LLMs for GRS.

Information access systems, such as Google News or YouTube, increasingly employ algorithms to rank diverse content such as music, recipes, and news articles. Acknowledging the influential role of these algorithms as gatekeepers to online content, the research community is increasingly exploring ‘beyond-accuracy’ metrics. However, deciding what norms and values are relevant and should be prioritized when designing and evaluating information access systems is a challenging task. This tutorial aims to cultivate normative thinking and decision-making in the design and evaluation of information access systems. The tutorial comprises two key components. The first part involves a lecture on the foundational principles of normative thinking, emphasizing the importance of reflecting on the desired state of a system rather than its current state. The second part is an interactive session where participants engage in group discussions, applying normative thinking to a specific use case. Participants analyze the system’s usage, stakeholders, and relevant norms and values and address potential conflicts between stakeholders and/or values. Through a point-allocation exercise, participants represent stakeholders and advocate for specific values, fostering a deeper understanding of normative decision-making in the context of information access systems.

Recommender systems are among the most widely used applications of artificial intelligence. Their use can have far-reaching consequences for users, stakeholders, and society at large. In this second edition of the NORMalize workshop, we once again seek to advance the research agenda of normative thinking, considering the norms and values that underpin recommender systems, as well as to introduce the concept to a broader audience. We aim to bring together a growing community of researchers and practitioners across disciplines who want to think about the norms and values that should be considered in the design and evaluation of recommender systems, and to further educate them on how to reflect on, prioritise, and operationalise such norms and values. NORMalize 2024 is a half-day workshop consisting of a combination of paper presentations and an interactive session, building upon its successful full-day run last year at RecSys’23.

When people use web search engines to find information on debated topics, the search results they encounter can influence opinion formation and practical decision-making with potentially far-reaching consequences for the individual and society. However, current web search engines lack support for information-seeking strategies that enable responsible opinion formation, e.g., by mitigating confirmation bias and motivating engagement with diverse viewpoints. We conducted two preregistered user studies to test the benefits and risks of an intervention aimed at confirmation bias mitigation. In the first study, we tested the effect of warning labels, warning of the risk of confirmation bias, combined with obfuscations, hiding selected search results per default. We observed that obfuscations with warning labels effectively reduce engagement with search results. These initial findings did not allow conclusions about the extent to which the reduced engagement was caused by the warning label (reflective nudging element) versus the obfuscation (automatic nudging element). If obfuscation was the primary cause, this would raise concerns about harming user autonomy. We thus conducted a follow-up study to test the effect of warning labels and obfuscations separately. According to our findings, obfuscations run the risk of manipulating behavior instead of guiding it, while warning labels without obfuscations (purely reflective) do not exhaust processing capacities but encourage users to actively choose to decrease engagement with attitude-confirming search results. Therefore, given the risks and unclear benefits of obfuscations and potentially other automatic nudging elements to guide engagement with information, we call for prioritizing interventions that aim to enhance human cognitive skills and agency instead.

One way to help users navigate debated topics online is to apply stance detection in web search. Automatically identifying whether search results are against, neutral, or in favor could facilitate diversification efforts and support interventions that aim to mitigate cognitive biases. To be truly useful in this context, however, stance detection models not only need to make accurate (cross-topic) predictions but also be sufficiently explainable to users when applied to search results – an issue that is currently unclear. This paper presents a study into the feasibility of using current stance detection approaches to assist users in their web search on debated topics. We train and evaluate 10 stance detection models using a stance-annotated data set of 1204 search results. In a preregistered user study (N = 291), we then investigate the quality of stance detection explanations created using different explainability methods and explanation visualization techniques. The models we implement predict stances of search results across topics with satisfying quality (i.e., similar to the state-of-the-art for other data types). However, our results reveal stark differences in explanation quality (i.e., as measured by users’ ability to simulate model predictions and their attitudes towards the explanations) between different models and explainability methods. A qualitative analysis of textual user feedback further reveals potential application areas, user concerns, and improvement suggestions for such explanations. Our findings have important implications for the development of user-centered solutions surrounding web search on debated topics.

Recommender systems are among the most widely used applications of artificial intelligence. Since they are so widely used, it is important that we, as practitioners and researchers, think about the impact these systems may have on users, society, and other stakeholders. To that effect, the NORMalize workshop seeks to introduce normative thinking, to consider the norms and values that underpin recommender systems in the recommender systems community. The objective of NORMalize is to bring together a growing community of researchers and practitioners across disciplines who want to think about the norms and values that should be considered in the design and evaluation of recommender systems; and further educate them on how to reflect on, prioritise, and operationalise such norms and values. NORMalize offers a comprehensive program designed to cater to both the norm-curious and the norm-active. The morning session is on-site and features a lecture on normative thinking and an interactive workshop. The afternoon is a hybrid program focused on the dissemination of results. NORMalize publishes proceedings, as well as a technical report that summarises the outcomes of the interactive morning session.

Increasingly taking place in online spaces, modern political conversations are typically perceived to be unproductively affirming---siloed in so called "echo chambers" of exclusively like-minded discussants. Yet, to date we lack sufficient means to measure viewpoint diversity in conversations. To this end, in this paper, we operationalize two viewpoint metrics proposed for recommender systems and adapt them to the context of social media conversations. This is the first study to apply these two metrics (Representation and Fragmentation) to real world data and to consider the implications for online conversations specifically. We apply these measures to two topics---daylight savings time (DST), which serves as a control, and the more politically polarized topic of immigration. We find that the diversity scores for both Fragmentation and Representation are lower for immigration than for DST. Further, we find that while pro-immigrant views receive consistent pushback on the platform, anti-immigrant views largely operate within echo chambers. We observe less severe yet similar patterns for DST. Taken together, Representation and Fragmentation paint a meaningful and important new picture of viewpoint diversity.

Research in the area of human information interaction (HII) typically represents viewpoints on debated topics in a binary fashion, as either against or in favor of a given topic (e.g., the feminist movement). This simple taxonomy, however, greatly reduces the latent richness of viewpoints and thereby limits the potential of research and practical applications in this field. Work in the communication sciences has already demonstrated that viewpoints can be represented in much more comprehensive ways, which could enable a deeper understanding of users’ interactions with debated topics online. For instance, a viewpoint’s stance usually has a degree of strength (e.g., mild or strong), and, even if two viewpoints support or oppose something to the same degree, they may use different logics of evaluation (i.e., underlying reasons). In this paper, we draw from communication science practice to propose a novel, two-dimensional way of representing viewpoints that incorporates a viewpoint’s stance degree as well as its logic of evaluation. We show in a case study of tweets on debated topics how our proposed viewpoint label can be obtained via crowdsourcing with acceptable reliability. By analyzing the resulting data set and conducting a user study, we further show that the two-dimensional viewpoint representation we propose allows for more meaningful analyses and diversification interventions compared to current approaches. Finally, we discuss what this novel viewpoint label implies for HII research and how obtaining it may be made cheaper in the future.

Recent research has demonstrated that cognitive biases such as the confirmation bias or the anchoring effect can negatively affect the quality of crowdsourced data. In practice, however, such biases go unnoticed unless specifically assessed or controlled for. Task requesters need to ensure that task workflow and design choices do not trigger workers’ cognitive biases. Moreover, to facilitate the reuse of crowdsourced data collections, practitioners can benefit from understanding whether and which cognitive biases may be associated with the data. To this end, we propose a 12-item checklist adapted from business psychology to combat cognitive biases in crowdsourcing. We demonstrate the practical application of this checklist in a case study on viewpoint annotations for search results. Through a retrospective analysis of relevant crowdsourcing research that has been published at HCOMP in 2018, 2019, and 2020, we show that cognitive biases may often affect crowd workers but are typically not considered as potential sources of poor data quality. The checklist we propose is a practical tool that requesters can use to improve their task designs and appropriately describe potential limitations of collected data. It contributes to a body of efforts towards making human-labeled data more reliable and reusable.

In web search on debated topics, algorithmic and cognitive biases strongly influence how users consume and process information. Recent research has shown that this can lead to a search engine manipulation effect (SEME): when search result rankings are biased towards a particular viewpoint, users tend to adopt this favored viewpoint. To better understand the mechanisms underlying SEME, we present a pre-registered, 5 x 3 factorial user study investigating whether order effects (i.e., users adopting the viewpoint pertaining to higher-ranked documents) can cause SEME. For five different debated topics, we evaluated attitude change after exposing participants with mild pre-existing attitudes to search results that were overall viewpoint-balanced but reflected one of three levels of algorithmic ranking bias. We found that attitude change did not differ across levels of ranking bias and did not vary based on individual user differences. Our results thus suggest that order effects may not be an underlying mechanism of SEME. Exploratory analyses lend support to the presence of exposure effects (i.e., users adopting the majority viewpoint among the results they examine) as a contributing factor to users' attitude change. We discuss how our findings can inform the design of user bias mitigation strategies.

For group recommendations, one objective is to recommend an ordered set of items, a top-N, to a group such that each individual recommendation is relevant for everyone. A common way to do this is to select items on which the group can agree, using so-called ‘aggregation strategies’. One weakness of these aggregation strategies is that they select items independently of each other. They therefore cannot guarantee properties such as fairness, that apply to the set of recommendations as a whole.

Previous research has found that enabling users to control the recommendation process increases user satisfaction. However, providing additional controls also increases cognitive load, and different users have different needs for control. Therefore, in this study, we investigate the effect of two personal characteristics: musical sophistication and visual memory capacity. We designed a visual user interface, on top of a commercial music recommender, with different controls: interactions with recommendations (i.e., the output of a recommender system), the user profile (i.e., the top listened songs), and algorithm parameters (i.e., weights in an algorithm). We created eight experimental settings with combinations of these three user controls and conducted a between-subjects study (N=240), to explore the effect on cognitive load and recommendation acceptance for different personal characteristics. We found that controlling recommendations is the most favorable single control element. In addition, controlling user profile and algorithm parameters was the most beneficial setting with multiple controls. Moreover, the participants with high musical sophistication perceived recommendations to be of higher quality, which in turn lead to higher recommendation acceptance. However, we found no effect of visual working memory on either cognitive load or recommendation acceptance. This work contributes an understanding of how to design control that hits the sweet spot between the perceived quality of recommendations and acceptable cognitive load.

As intelligent interactive systems, recommender systems focus on determining predictions that fit the wishes and needs of users. Still, a large majority of recommender systems research focuses on accuracy criteria and much less attention is paid to how users interact with the system, and in which way the user interface has an influence on the selection behavior of the users. Consequently, it is important to look beyond algorithms. The main goals of the IntRS workshop are to analyze the impact of user interfaces and interaction design, and to explore human interaction with recommender systems from a human decision making perspective. Methodologies for evaluating these aspects are also within the scope of the workshop.

As intelligent interactive systems, recommender systems focus on determining predictions that fit the wishes and needs of users. Still, a large majority of recommender systems research focuses on accuracy criteria and much less attention is paid to how users interact with the system, and in which way the user interface has an influence on the selection behavior of the users. Consequently, it is important to look beyond algorithms. The main goals of the IntRS workshop are to analyze the impact of user interfaces and interaction design, and to explore human interaction with recommender systems. Methodologies for evaluating these aspects are also within the scope of the workshop.

As an interactive intelligent system, recommender systems are developed to give predictions that match users preferences. Since the emergence of recommender systems, a large majority of research focuses on objective accuracy criteria and less attention has been paid to how users interact with the system and the efficacy of interface designs from the end-user perspective. The field has reached a point where it is ready to look beyond algorithms, into users interactions, decision making processes and overall experience. Accordingly, the goals of this workshop (Int[email protected]) are to explore the human aspects of recommender systems, with a particular focus on the impact of interfaces and interaction design on decision-making and user experiences with recommender systems, and to explore methodologies to evaluate these human aspects of the recommendation process that go beyond traditional automated approaches.

No abstract available.

A common approach to designing Recommender Systems (RS) consists of asking users to explicitly rate items in order to collect feedback about their preferences. However, users have been shown to be inconsistent and to introduce a non-negligible amount of natural noise in their ratings that affects the accuracy of the predictions. In this paper, we present a novel approach to improve RS accuracy by reducing the natural noise in the input data via a preprocessing step. In order to quantitatively understand the impact of natural noise, we first analyze the response of common recommendation algorithms to this noise. Next, we propose a novel algorithm to denoise existing datasets by means of re-rating: i.e. by asking users to rate previously rated items again. This denoising step yields very significant accuracy improvements. However, re-rating all items in the original dataset is unpractical. Therefore, we study the accuracy gains obtained when re-rating only some of the ratings.In particular, we propose two partial denoising strategies: data and user-dependent denoising. Finally, we compare the value of adding a rating of an unseen item vs. re-rating an item. We conclude with a proposal for RS to improve the quality of their user data and hence their accuracy: asking users to re-rate items might, in some circumstances, be more beneficial than asking users to rate unseen items.

This thesis focuses on explanations of recommendations. Explanations can have many advantages, from inspiring user trust to helping users make good decisions. We have identified seven different aims of explanations, and in this thesis we will consider how explanations can be optimized for some of these aims. We will consider both an explanation's content and its presentation. As a domain, we are currently investigating explanations for a movie recommender, and developing a prototype system. This paper summarizes the goals of the thesis, the methodology we are using, the work done so far and our intended future work.

This paper characterizes general properties of useful, or Effective, explanations of recommendations. It describes a methodology based on focus groups, in which we elicit what helps moviegoers decide whether or not they would like a movie. Our results highlight the importance of personalizing explanations to the individual user, as well as considering the source of recommendations, user mood, the effects of group viewing, and the effect of explanations on user expectations.