RU
R. Ungruh
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
11 records found
1
When Preference Is Not Enough
Why Recommender Systems Require Human-Aware Evaluation for Children
Children regularly interact with recommender systems, yet little is known about whether the suggestions they encounter fit them. Traditional accuracy-based evaluation accounts for user preference, providing an incomplete picture of how well recommenders serve young users. With that in mind, we adopt a human-centric, specifically child-centric, evaluation perspective to empirically examine whether recommender systems address children's needs. Focusing on content maturity as a key dimension influencing what is considered fitting as per developmental needs, we probe a range of recommender algorithms on whether their suggestions align with children's developmental maturity and how closely these suggestions reflect the content maturity of previously consumed items. Our analysis showcases that traditional evaluation paradigms fail to uncover dynamics that affect alignment with children's needs.
...
Children regularly interact with recommender systems, yet little is known about whether the suggestions they encounter fit them. Traditional accuracy-based evaluation accounts for user preference, providing an incomplete picture of how well recommenders serve young users. With that in mind, we adopt a human-centric, specifically child-centric, evaluation perspective to empirically examine whether recommender systems address children's needs. Focusing on content maturity as a key dimension influencing what is considered fitting as per developmental needs, we probe a range of recommender algorithms on whether their suggestions align with children's developmental maturity and how closely these suggestions reflect the content maturity of previously consumed items. Our analysis showcases that traditional evaluation paradigms fail to uncover dynamics that affect alignment with children's needs.
Recommender systems increasingly shape children's online experiences, yet the risks they might pose remain poorly understood. Formal risk assessments are rare, overlook children's experiences, treat them as a homogeneous group, and often consider risks on a population-wide level. To provide a more nuanced child-centered perspective on risks associated with recommenders, we conduct a comprehensive risk assessment that examines multiple risks simultaneously and captures individual-level exposure rather than solely aggregated metrics. We also explore behavioral characteristics that make some children more susceptible to risk exposure to grasp how individual differences shape the manifestation of risks. Our results highlight the need for risk assessments that recognize heterogeneity among children and understand when risks become systemic' through individualized recommendation patterns. These insights spark the discussion about child-aware recommenders and ongoing regulatory efforts for evidence-based accountability.
...
Recommender systems increasingly shape children's online experiences, yet the risks they might pose remain poorly understood. Formal risk assessments are rare, overlook children's experiences, treat them as a homogeneous group, and often consider risks on a population-wide level. To provide a more nuanced child-centered perspective on risks associated with recommenders, we conduct a comprehensive risk assessment that examines multiple risks simultaneously and captures individual-level exposure rather than solely aggregated metrics. We also explore behavioral characteristics that make some children more susceptible to risk exposure to grasp how individual differences shape the manifestation of risks. Our results highlight the need for risk assessments that recognize heterogeneity among children and understand when risks become systemic' through individualized recommendation patterns. These insights spark the discussion about child-aware recommenders and ongoing regulatory efforts for evidence-based accountability.
Blurred Lines
Understanding the Fit of Song Lyrics in Music Catalogs That Can Reach Children Through Recommendations
Recommender systems on popular online platforms expose impressionable and easily influenced younger listeners to varied content, making it crucial to reflect on the songs children can encounter due to their interactions with recommender systems. To set a foundation, we analyze the lyrics of a catalog comprised of ∼30,000 songs to gauge their suitability to children. Our multi-perspective exploration reveals a high prevalence of inappropriate lyrics in music commonly heard by children. This highlights the need for further explorations pertaining online platforms and their recommender systems that curate and ultimately present items from catalogs such as the ones we examined, highlighting the potential negative impact of such lyrics on their behavior and personality by promoting harmful language or biases. Informed by our findings, we outline research directions for the information retrieval community to consider when designing, evaluating, and deploying algorithms that serve diverse audiences.
...
Recommender systems on popular online platforms expose impressionable and easily influenced younger listeners to varied content, making it crucial to reflect on the songs children can encounter due to their interactions with recommender systems. To set a foundation, we analyze the lyrics of a catalog comprised of ∼30,000 songs to gauge their suitability to children. Our multi-perspective exploration reveals a high prevalence of inappropriate lyrics in music commonly heard by children. This highlights the need for further explorations pertaining online platforms and their recommender systems that curate and ultimately present items from catalogs such as the ones we examined, highlighting the potential negative impact of such lyrics on their behavior and personality by promoting harmful language or biases. Informed by our findings, we outline research directions for the information retrieval community to consider when designing, evaluating, and deploying algorithms that serve diverse audiences.
Breaking the loop
A human-centred vision for information retrieval systems that serve all users
Journal article
(2026)
-
Angelo Geninatti Cossatin, H. Chakrabarti, Emiliana Murgia, Simone Ooms, Diletta Micol Tobia, R. Ungruh, Federica Cena, Stephane Aline Chaudron, Maria Soledad Pera, More Authors
Information retrieval (IR) systems have predominantly been designed for and evaluated in mainstream user populations, creating a self-reinforcing cycle of exclusion for underserved groups, e.g. differently-abled individuals, cultural minorities, those facing economic constraints, experiencing digital divides, language barriers, or those at age extremes. This opinion paper presents findings from a collaborative brainstorming session on underrepresented users in IR, identifying key challenges and pathways toward more inclusive design. Our analysis reveals a critical loop: lack of access leads to underrepresentation in datasets, which in turn leads to inadequate consideration in system design, perpetuating exclusion. We present guidelines for breaking this cycle through four interconnected dimensions: interface design grounded in participatory methods and universal accessibility, adaptive algorithms that respond to diverse capabilities and contexts, evaluation methodologies that position underserved users as active collaborators rather than passive subjects, and responsible AI integration guided by user-centred principles.
...
Information retrieval (IR) systems have predominantly been designed for and evaluated in mainstream user populations, creating a self-reinforcing cycle of exclusion for underserved groups, e.g. differently-abled individuals, cultural minorities, those facing economic constraints, experiencing digital divides, language barriers, or those at age extremes. This opinion paper presents findings from a collaborative brainstorming session on underrepresented users in IR, identifying key challenges and pathways toward more inclusive design. Our analysis reveals a critical loop: lack of access leads to underrepresentation in datasets, which in turn leads to inadequate consideration in system design, perpetuating exclusion. We present guidelines for breaking this cycle through four interconnected dimensions: interface design grounded in participatory methods and universal accessibility, adaptive algorithms that respond to diverse capabilities and contexts, evaluation methodologies that position underserved users as active collaborators rather than passive subjects, and responsible AI integration guided by user-centred principles.
Children form stereotypes by observing stereotypical expressions during childhood, influencing their future beliefs, attitudes, and behavior. These perceptions, often negative, can surface across the many online media platforms that children access, like streaming services and social media. Given that many of the items displayed on these platforms are commonly selected by recommendation algorithms (RAs), it becomes critical to investigate their role in suggesting items that could negatively impact this vulnerable population. We address this concern by conducting an empirical evaluation to gauge the presence of Gender, Race, and Religion stereotypes among the top-10 recommendations generated by a wide range of RAs across two well-known datasets in different domains: Movielens (movies) and GoodReads (books). Results analyses reveal that all RAs frequently recommend stereotypical items. Gender stereotypes are particularly prevalent, appearing in almost every recommendation list and emerging as the most common stereotype. Our results indicate that no algorithm has a consistent tendency towards recommending more stereotypical content; instead, high stereotype presence can be found across recommendation strategies. Outcomes from this work underscore the potential risks that RAs pose to children in perpetuating and reinforcing harmful stereotypes—this prompts reflections on their implications for the design and evaluation of recommender systems
...
Children form stereotypes by observing stereotypical expressions during childhood, influencing their future beliefs, attitudes, and behavior. These perceptions, often negative, can surface across the many online media platforms that children access, like streaming services and social media. Given that many of the items displayed on these platforms are commonly selected by recommendation algorithms (RAs), it becomes critical to investigate their role in suggesting items that could negatively impact this vulnerable population. We address this concern by conducting an empirical evaluation to gauge the presence of Gender, Race, and Religion stereotypes among the top-10 recommendations generated by a wide range of RAs across two well-known datasets in different domains: Movielens (movies) and GoodReads (books). Results analyses reveal that all RAs frequently recommend stereotypical items. Gender stereotypes are particularly prevalent, appearing in almost every recommendation list and emerging as the most common stereotype. Our results indicate that no algorithm has a consistent tendency towards recommending more stereotypical content; instead, high stereotype presence can be found across recommendation strategies. Outcomes from this work underscore the potential risks that RAs pose to children in perpetuating and reinforcing harmful stereotypes—this prompts reflections on their implications for the design and evaluation of recommender systems
Recommender systems research seldom considers children as a user group, and when it does, it is anchored on datasets where children are underrepresented, risking overlooking their interests, favoring those of the majority, i.e., mainstream users. Recently, Ungruh et al. demonstrated that children’s consumption patterns and preferences differ from those of mainstream users, resulting in inconsistent recommendation algorithm performance and behavior for this user group. These findings, however, are based on two datasets with a limited child user sample. We reproduce and replicate this study on a wider range of datasets in the movie, music, and book domains, uncovering interaction patterns and aspects of child-recommender interactions consistent across domains, as well as those specific to some user samples in the data. We also extend insights from the original study with popularity bias metrics, given the interpretation of results from the original study. With this reproduction and extension, we uncover consumption patterns and differences between age groups stemming from intrinsic differences between children and others, and those unique to specific datasets or domains.
...
Recommender systems research seldom considers children as a user group, and when it does, it is anchored on datasets where children are underrepresented, risking overlooking their interests, favoring those of the majority, i.e., mainstream users. Recently, Ungruh et al. demonstrated that children’s consumption patterns and preferences differ from those of mainstream users, resulting in inconsistent recommendation algorithm performance and behavior for this user group. These findings, however, are based on two datasets with a limited child user sample. We reproduce and replicate this study on a wider range of datasets in the movie, music, and book domains, uncovering interaction patterns and aspects of child-recommender interactions consistent across domains, as well as those specific to some user samples in the data. We also extend insights from the original study with popularity bias metrics, given the interpretation of results from the original study. With this reproduction and extension, we uncover consumption patterns and differences between age groups stemming from intrinsic differences between children and others, and those unique to specific datasets or domains.
Recommender Systems research continuously improves recommendation strategies to meet the needs of a wide range of users and other stakeholders. However, much of this research centers on the traditional, adult user, often overlooking underrepresented demographics. One such group is children, frequent users of platforms driven by recommender systems. Children differ from adults in preferences and can be particularly vulnerable to certain content, raising questions about the harm recommender systems may pose.
This PhD project advocates for child-aware recommender systems: systems that explicitly account for children as part of their users, recognizing their distinct needs, vulnerabilities, and rights. In pursuit of this goal, we investigate how well current recommender systems serve children, auditing algorithmic strategies from two complementary perspectives: The ‘traditional’ perspective focuses on whether recommendations align with children’s preferences. The perspective of ‘non-maleficence’ assesses suitability of content recommended, evaluating whether it respects children’s vulnerabilities to potentially harmful material. To do so, we audit current recommender systems according to both perspectives—not only in the short term, but also in the long term through simulation studies. Beyond auditing, we explore strategies and design directions for making recommender systems more responsible. Outcomes from this work should inform both academic and practitioner communities about the gaps in current systems and lay the groundwork for more equitable, safe, and meaningful recommendations for children. ...
This PhD project advocates for child-aware recommender systems: systems that explicitly account for children as part of their users, recognizing their distinct needs, vulnerabilities, and rights. In pursuit of this goal, we investigate how well current recommender systems serve children, auditing algorithmic strategies from two complementary perspectives: The ‘traditional’ perspective focuses on whether recommendations align with children’s preferences. The perspective of ‘non-maleficence’ assesses suitability of content recommended, evaluating whether it respects children’s vulnerabilities to potentially harmful material. To do so, we audit current recommender systems according to both perspectives—not only in the short term, but also in the long term through simulation studies. Beyond auditing, we explore strategies and design directions for making recommender systems more responsible. Outcomes from this work should inform both academic and practitioner communities about the gaps in current systems and lay the groundwork for more equitable, safe, and meaningful recommendations for children. ...
Recommender Systems research continuously improves recommendation strategies to meet the needs of a wide range of users and other stakeholders. However, much of this research centers on the traditional, adult user, often overlooking underrepresented demographics. One such group is children, frequent users of platforms driven by recommender systems. Children differ from adults in preferences and can be particularly vulnerable to certain content, raising questions about the harm recommender systems may pose.
This PhD project advocates for child-aware recommender systems: systems that explicitly account for children as part of their users, recognizing their distinct needs, vulnerabilities, and rights. In pursuit of this goal, we investigate how well current recommender systems serve children, auditing algorithmic strategies from two complementary perspectives: The ‘traditional’ perspective focuses on whether recommendations align with children’s preferences. The perspective of ‘non-maleficence’ assesses suitability of content recommended, evaluating whether it respects children’s vulnerabilities to potentially harmful material. To do so, we audit current recommender systems according to both perspectives—not only in the short term, but also in the long term through simulation studies. Beyond auditing, we explore strategies and design directions for making recommender systems more responsible. Outcomes from this work should inform both academic and practitioner communities about the gaps in current systems and lay the groundwork for more equitable, safe, and meaningful recommendations for children.
This PhD project advocates for child-aware recommender systems: systems that explicitly account for children as part of their users, recognizing their distinct needs, vulnerabilities, and rights. In pursuit of this goal, we investigate how well current recommender systems serve children, auditing algorithmic strategies from two complementary perspectives: The ‘traditional’ perspective focuses on whether recommendations align with children’s preferences. The perspective of ‘non-maleficence’ assesses suitability of content recommended, evaluating whether it respects children’s vulnerabilities to potentially harmful material. To do so, we audit current recommender systems according to both perspectives—not only in the short term, but also in the long term through simulation studies. Beyond auditing, we explore strategies and design directions for making recommender systems more responsible. Outcomes from this work should inform both academic and practitioner communities about the gaps in current systems and lay the groundwork for more equitable, safe, and meaningful recommendations for children.
Recommendation algorithms are often trained using data sources reflecting the interactions of a broad user base. As a result, the dominant preferences of the majority may overshadow those of other groups with unique interests. This is something performance analyses of recommendation algorithms typically fail to capture, prompting us to investigate how well recommendations align with preferences of the overall population but also specifically a “non-mainstream” user group: children—an audience frequently exposed to recommender systems but rarely prioritized. Using music and movie datasets, we examine the differences in genre preferences between Children and Mainstream Users. We then explore the degree to which (genre) consumption patterns of a mainstream group impact the recommendations classical algorithms offer children. Our findings highlight prominent differences in consumption patterns between Children and Mainstream Users; they also reflect that children’s recommendations are impacted by the preference of user groups with deviating consumption habits. Surprisingly, despite being under-represented, children do not necessarily receive poorer recommendations. Further, our results demonstrate that tailoring training specifically to children does not always enhance personalization for them. These findings prompt reflections and discussion on how recommender systems can better meet the needs of understudied user groups.
...
Recommendation algorithms are often trained using data sources reflecting the interactions of a broad user base. As a result, the dominant preferences of the majority may overshadow those of other groups with unique interests. This is something performance analyses of recommendation algorithms typically fail to capture, prompting us to investigate how well recommendations align with preferences of the overall population but also specifically a “non-mainstream” user group: children—an audience frequently exposed to recommender systems but rarely prioritized. Using music and movie datasets, we examine the differences in genre preferences between Children and Mainstream Users. We then explore the degree to which (genre) consumption patterns of a mainstream group impact the recommendations classical algorithms offer children. Our findings highlight prominent differences in consumption patterns between Children and Mainstream Users; they also reflect that children’s recommendations are impacted by the preference of user groups with deviating consumption habits. Surprisingly, despite being under-represented, children do not necessarily receive poorer recommendations. Further, our results demonstrate that tailoring training specifically to children does not always enhance personalization for them. These findings prompt reflections and discussion on how recommender systems can better meet the needs of understudied user groups.
From Previous Plays to Long-Term Tastes
Exploring the Long-term Reliability of Recommender Systems Simulations for Children
Studying the interplay of children and recommender systems (RS) is ethically and practically challenging, making simulation a promising alternative for exploration. However, recent simulation approaches that aim to model natural user-RS interactions typically rely on behavioral data and assume that user preferences remain consistent over time—an assumption that may not hold for children who undergo continuous developmental changes. With that in mind, we explore the extent to which simulations based on historical data can meaningfully reflect children’s long-term consumption patterns. We do this via a simulation study using real-world data in which user behavior is modeled from observed listening preferences. Specifically, we probe whether simulation mirrors user preferences over time by comparing with organic (i.e., real) consumption patterns. Our findings offer a critical reflection on the reliability of simulation-based RS research for children and question the reliability of using behavioral assumptions to model users.
...
Studying the interplay of children and recommender systems (RS) is ethically and practically challenging, making simulation a promising alternative for exploration. However, recent simulation approaches that aim to model natural user-RS interactions typically rely on behavioral data and assume that user preferences remain consistent over time—an assumption that may not hold for children who undergo continuous developmental changes. With that in mind, we explore the extent to which simulations based on historical data can meaningfully reflect children’s long-term consumption patterns. We do this via a simulation study using real-world data in which user behavior is modeled from observed listening preferences. Specifically, we probe whether simulation mirrors user preferences over time by comparing with organic (i.e., real) consumption patterns. Our findings offer a critical reflection on the reliability of simulation-based RS research for children and question the reliability of using behavioral assumptions to model users.
Simulation is widely used in recommender systems research to study algorithm behavior and its impact on users. A common strategy involves adopting a universal choice model to represent users, assuming all follow the same consumption patterns. This one-size-fits-all approach overlooks the diversity in user preferences and decision-making patterns. In this work, we scrutinize whether this universal view fails to account for unique user behavior, thus harming realism and reliability of simulation outcomes. We conduct multiple simulations with various recommendation algorithms and choice models in the movie domain, comparing outcomes to users’ organic consumption patterns. Further, we evaluate whether a holistic model that captures users’ differences in behavior would better reflect a wide user base. Our findings highlight the limitations of using a naive, universal choice model and emphasize the need for more nuanced, user-specific approaches to make contributions from simulation studies more reflective of real-world effects.
...
Simulation is widely used in recommender systems research to study algorithm behavior and its impact on users. A common strategy involves adopting a universal choice model to represent users, assuming all follow the same consumption patterns. This one-size-fits-all approach overlooks the diversity in user preferences and decision-making patterns. In this work, we scrutinize whether this universal view fails to account for unique user behavior, thus harming realism and reliability of simulation outcomes. We conduct multiple simulations with various recommendation algorithms and choice models in the movie domain, comparing outcomes to users’ organic consumption patterns. Further, we evaluate whether a holistic model that captures users’ differences in behavior would better reflect a wide user base. Our findings highlight the limitations of using a naive, universal choice model and emphasize the need for more nuanced, user-specific approaches to make contributions from simulation studies more reflective of real-world effects.
Conference paper
(2024)
-
Robin Ungruh, Karlijn Dinnissen, Anja Volk, Maria Soledad Pera, Hanna Hauptmann
Popularity bias is a prominent phenomenon in recommender systems (RS), especially in the music domain. Although popularity bias mitigation techniques are known to enhance the fairness of RS while maintaining their high performance, there is a lack of understanding regarding users’ actual perception of the suggested music. To address this gap, we conducted a user study (n=40) exploring user satisfaction and perception of personalized music recommendations generated by algorithms that explicitly mitigate popularity bias. Specifically, we investigate item-centered and user-centered bias mitigation techniques, aiming to ensure fairness for artists or users, respectively. Results show that neither mitigation technique harms the users’ satisfaction with the recommendation lists despite promoting underrepresented items. However, the item-centered mitigation technique impacts user perception; by promoting less popular items, it reduces users’ familiarity with the items. Lower familiarity evokes discovery—the feeling that the recommendations enrich the user’s taste. We demonstrate that this can ultimately lead to higher satisfaction, highlighting the potential of less-popular recommendations to improve the user experience.
...
Popularity bias is a prominent phenomenon in recommender systems (RS), especially in the music domain. Although popularity bias mitigation techniques are known to enhance the fairness of RS while maintaining their high performance, there is a lack of understanding regarding users’ actual perception of the suggested music. To address this gap, we conducted a user study (n=40) exploring user satisfaction and perception of personalized music recommendations generated by algorithms that explicitly mitigate popularity bias. Specifically, we investigate item-centered and user-centered bias mitigation techniques, aiming to ensure fairness for artists or users, respectively. Results show that neither mitigation technique harms the users’ satisfaction with the recommendation lists despite promoting underrepresented items. However, the item-centered mitigation technique impacts user perception; by promoting less popular items, it reduces users’ familiarity with the items. Lower familiarity evokes discovery—the feeling that the recommendations enrich the user’s taste. We demonstrate that this can ultimately lead to higher satisfaction, highlighting the potential of less-popular recommendations to improve the user experience.