This workshop aims to empower participants to design and audit recommendation systems that prioritize user trust and safety. Atten- dees will explore best practices and innovations in fairness, explain- ability, content safety, algorithm transparency, and the societal im- pacts of recommendation systems. The workshop will address tech- nical and ethical challenges, offering practical insights into risk mit- igation, social concerns, and technical innovations. Participants will leave equipped with actionable frameworks, methodologies, and tools to build responsible systems aligned with societal values and social good. Key activities include keynotes, paper presentations, panel discussions, and interactive sessions, covering topics such as transparency, security, fairness, robustness, and participatory AI. The website of the conference is https://sites.google.com/view/t-rrs.

Algorithmic fairness in the context of personalized recommendation presents significantly different challenges to those commonly encountered in classification tasks. Researchers studying classification have generally considered fairness to be a matter of achieving equality of outcomes (or some other metric) between a protected and unprotected group and built algorithmic interventions on this basis. We argue that fairness in real-world application settings in general, and especially in the context of personalized recommendation, is much more complex and multi-faceted, requiring a more general approach. To address the fundamental problem of fairness in the presence of multiple stakeholders, with different definitions of fairness, we propose the Social Choice for Recommendation Under Fairness–Dynamic architecture, which formalizes multistakeholder fairness in recommender systems as a two-stage social choice problem. In particular, we express recommendation fairness as a combination of an allocation and an aggregation problem, which integrate both fairness concerns and personalized recommendation provisions, and derive new recommendation techniques based on this formulation. We demonstrate the ability of our framework to dynamically incorporate multiple fairness concerns using both real-world and synthetic datasets.

Algorithmic fairness in recommender systems requires close attention to the needs of a diverse set of stakeholders that may have competing interests. Previous work in this area has often been limited by fixed, single-objective definitions of fairness, built into algorithms or optimization criteria that are applied to a single fairness dimension or, at most, applied identically across dimensions. These narrow conceptualizations limit the ability to adapt fairness-aware solutions to the wide range of stakeholder needs and fairness definitions that arise in practice. Our work approaches recommendation fairness from the standpoint of computational social choice, using a multi-agent framework. In this paper, we explore the properties of different social choice mechanisms and demonstrate the successful integration of multiple, heterogeneous fairness definitions across multiple data sets.

The Platform for OPen Recommendation and Online eXperimentation (POPROX) is a new resource to allow RecSys researchers to conduct online user research without having to develop all of the necessary infrastructure and recruit users. Our first domain is personalized news recommendations – POPROX 1.0 provides a daily newsletter (with content from the Associated Press) to users who have already consented to participate in research, along with interfaces and protocols to support researchers in conducting studies that assign subsets of users to various experimental algorithms and/or interfaces.

Despite increasing reliance on personalization in digital platforms, many algorithms that curate content or information for users have been met with resistance. When users feel dissatisfied or harmed by recommendations, this can lead users to hate, or feel negatively towards these personalized systems. Algorithmic hate detrimentally impacts both users and the system, and can result in various forms of algorithmic harm, or in extreme cases can lead to public protests against “the algorithm” in question. In this work, we summarize some of the most common causes of algorithmic hate and their negative consequences through various case studies of personalized recommender systems. We explore promising future directions for the RecSys research community that could help alleviate algorithmic hate and improve the relationship between recommender systems and their users.

The 5th Workshop on Responsible Recommendation (FAccTRec 2022) was held in conjunction with the 16th ACM Conference on Recommender Systems on September, 2022 at Seattle, USA, in a hybrid format. This workshop brought together researchers and practitioners to discuss several topics under the banner of social responsibility in recommender systems: fairness, accountability, transparency, privacy, and other ethical and social concerns. It served to advance research and discussion of these topics in the recommender systems space, and incubate ideas for future development and refinement.

Fairness is a critical system-level objective in recommender systems that has been the subject of extensive recent research. A specific form of fairness is supplier exposure fairness, where the objective is to ensure equitable coverage of items across all suppliers in recommendations provided to users. This is especially important in multistakeholder recommendation scenarios where it may be important to optimize utilities not just for the end user but also for other stakeholders such as item sellers or producers who desire a fair representation of their items. This type of supplier fairness is sometimes accomplished by attempting to increase aggregate diversity to mitigate popularity bias and to improve the coverage of long-tail items in recommendations. In this article, we introduce FairMatch, a general graph-based algorithm that works as a post-processing approach after recommendation generation to improve exposure fairness for items and suppliers. The algorithm iteratively adds high-quality items that have low visibility or items from suppliers with low exposure to the users’ final recommendation lists. A comprehensive set of experiments on two datasets and comparison with state-of-the-art baselines show that FairMatch, although it significantly improves exposure fairness and aggregate diversity, maintains an acceptable level of relevance of the recommendations.

Recommender systems are complex. They integrate the individual needs of users with the characteristics of particular domains of application which may span items from large and potentially heterogeneous collections. Extensive experimentation is required to understand the multidimensional properties of recommendation algorithms and the fit between algorithm and application. librec-auto is a tool that automates many aspects of off-line batch recommender system experimentation. It has a large library of state-of-the-art and historical recommendation algorithms and a wide variety of evaluation metrics. It further supports the study of diversity and fairness in recommendation through the integration of re-ranking algorithms and fairness-aware metrics. It supports declarative configuration for reproducible experiment management and supports multiple forms of hyper-parameter optimization.

There is significant interest lately in using synthetic data and simulation infrastructures for various types of recommender systems research. However, there are not currently any clear best practices around how best to apply these methods. We proposed a workshop to bring together researchers and practitioners interested in simulating recommender systems and their data to discuss the state of the art of such research and the pressing open methodological questions. The workshop resulted in a report authored by the participants that documents currently-known best practices on which the group has consensus and lays out an agenda for further research over the next 3–5 years to fill in places where we currently lack the information needed to make methodological recommendations.

It is well known that explicit user ratings in recommender systems are biased toward high ratings and that users differ significantly in their usage of the rating scale. Implementers usually compensate for these issues through rating normalization or the inclusion of a user bias term in factorization models. However, these methods adjust only for the central tendency of users’ distributions. In this work, we demonstrate that a lack of flatness in rating distributions is negatively correlated with recommendation performance. We propose a rating transformation model that compensates for skew in the rating distribution as well as its central tendency by converting ratings into percentile values as a pre-processing step before recommendation generation. This transformation flattens the rating distribution, better compensates for differences in rating distributions, and improves recommendation performance. We also show that a smoothed version of this transformation can yield more intuitive results for users with very narrow rating distributions. A comprehensive set of experiments, with state-of-the-art recommendation algorithms in four real-world datasets, show improved ranking performance for these percentile transformations.

Recommendation algorithms are known to suffer from popularity bias; a few popular items are recommended frequently while the majority of other items are ignored. These recommendations are then consumed by the users, their reaction will be logged and added to the system: what is generally known as a feedback loop. In this paper, we propose a method for simulating the users interaction with the recommenders in an offline setting and study the impact of feedback loop on the popularity bias amplification of several recommendation algorithms. We then show how this bias amplification leads to several other problems such as declining the aggregate diversity, shifting the representation of users' taste over time and also homogenization of the users. In particular, we show that the impact of feedback loop is generally stronger for the users who belong to the minority group.

Recently there has been a growing interest in fairness-aware recommender systems including fairness in providing consistent performance across different users or groups of users. A recommender system could be considered unfair if the recommendations do not fairly represent the tastes of a certain group of users while other groups receive recommendations that are consistent with their preferences. In this paper, we use a metric called miscalibration for measuring how a recommendation algorithm is responsive to users’ true preferences and we consider how various algorithms may result in different degrees of miscalibration for different users. In particular, we conjecture that popularity bias which is a well-known phenomenon in recommendation is one important factor leading to miscalibration in recommendation. Our experimental results using two real-world datasets show that there is a connection between how different user groups are affected by algorithmic popularity bias and their level of interest in popular items. Moreover, we show that the more a group is affected by the algorithmic popularity bias, the more their recommendations are miscalibrated.

Comparative experimentation is important for studying reproducibility in recommender systems. This is particularly true in areas without well-established methodologies, such as fairness-aware recommendation. In this paper, we describe fairness-aware enhancements to our recommender systems experimentation tool librec-auto. These enhancements include metrics for various classes of fairness definitions, extension of the experimental model to support result re-ranking and a library of associated re-ranking algorithms, and additional support for experiment automation and reporting. The associated demo will help attendees move quickly to configuring and running their own experiments with librec-auto.

In research practice, recommender systems are typically evaluated on their ability to provide items that satisfy the needs and interests of the end user. However, in many recommendation domains, the user for whom recommendations are generated is not the only stakeholder in the recommendation outcome. For example, fairness and balance across stakeholders is important in some recommendation applications; achieving a goal such as promoting new sellers in a marketplace might be important in others. Such multistakeholder environments present unique challenges for recommender system design and evaluation, and these challenges were the focus of this workshop.

Fairness and related concerns have become of increasing importance in a variety of AI and machine learning contexts. They are also highly relevant to recommender systems and related problems such as information retrieval, as evidenced by the growing literature in RecSys, FAT*, SIGIR, and special sessions such as the FATREC and FACTS-IR workshops and the Fairness track at TREC 2019; however, translating algorithmic fairness constructs from classification, scoring, and even many ranking settings into recommendation and other information access scenarios is not a straightforward task. This tutorial will help orient RecSys researchers to algorithmic fairness, understand how concepts do and do not translate from other settings, and provide an introduction to the growing literature on this topic.

Microlending can lead to improved access to capital in impoverished countries. Recommender systems could be used in microlending to provide efficient and personalized service to lenders. However, increasing concerns about discrimination in machine learning hinder the application of recommender systems to the microfinance industry. Most previous recommender systems focus on pure personalization, with fairness issue largely ignored. A desirable fairness property in microlending is to give borrowers from different demographic groups a fair chance of being recommended, as stated by Kiva. To achieve this goal, we propose a Fairness-Aware Re-ranking (FAR) algorithm to balance ranking quality and borrower-side fairness. Furthermore, we take into consideration that lenders may differ in their receptivity to the diversification of recommended loans, and develop a Personalized Fairness-Aware Re-ranking (PFAR) algorithm. Experiments on a real-world dataset from Kiva.org show that our re-ranking algorithm can significantly promote fairness with little sacrifice in accuracy, and be attentive to individual lender preference on loan diversity.

Fairness and related concerns have become of increasing importance in a variety of AI and machine learning contexts. They are also highly relevant to information retrieval and related problems such as recommendation, as evidenced by the growing literature in SIGIR, FAT*, RecSys, and special sessions such as the FATREC workshop and the Fairness track at TREC 2019; however, translating algorithmic fairness constructs from classification, scoring, and even many ranking settings into information retrieval and recommendation scenarios is not a straightforward task. This tutorial will help to orient IR researchers to algorithmic fairness, understand how concepts do and do not translate from other settings, and provide an introduction to the growing literature on this topic.

Recommender systems research often requires the creation and execution of large numbers of algorithmic experiments to determine the sensitivity of results to the values of various hyperparameters. Existing recommender systems platforms fail to provide a basis for systematic experimentation of this type. In this paper, we describe librec-auto, a wrapper for the well-known LibRec library, which provides an environment that supports automated experimentation.

Recommender systems are typically designed to optimize the utility of the end user. In many settings, however, the end user is not the only stakeholder and this exclusive focus may produce unsatisfactory results for other stakeholders. One such setting is found in multisided platforms, which bring together buyers and sellers. In such platforms, it may be necessary to jointly optimize the value for both buyers and sellers. This paper proposes a constraint-based integer programming optimization model, in which different sets of constraints are used to reflect the goals of the different stakeholders. This model is applied as a post-processing step, so it can easily be added onto an existing recommendation system to make it multi-stakeholder aware. For computational tractability with larger data sets, we reformulate the integer problem using the Lagrangian dual and use subgradient optimization. In experiments with two data sets, we evaluate empirically the interaction between the utilities of buyers and sellers and show that our approximation can achieve good upper and lower bounds in practical situations.

Many recommendation algorithms suffer from popularity bias in their output: popular items are recommended frequently and less popular ones rarely, if at all. However, less popular, long-tail items are precisely those that are often desirable recommendations. In this paper, we introduce a flexible regularization-based framework to enhance the long-tail coverage of recommendation lists in a learning-to-rank algorithm. We show that regularization provides a tunable mechanism for controlling the trade-off between accuracy and coverage. Moreover, the experimental results using two data sets show that it is possible to improve coverage of long tail items without substantial loss of ranking performance.

In this paper, we summarize VAMS 2017 - a workshop on value-aware and multistakeholder recommendation co-located with RecSys 2017. The workshop encouraged forward-thinking papers in this new area of recommender systems research and obtained a diverse set of responses ranging from application results to research overviews.

Recommender systems are key components in information-seeking contexts where personalization is sought. However, the dominant framework for recommendation is essentially two dimensional, with the interaction between users and items characterized by a single relation. In many cases, such as social networks, users and items are joined in a complex web of relations, not readily reduced to a single value. Recent multirelational approaches to recommendation focus on the direct, proximal relations in which users and items may participate. Our approach uses the framework of complex heterogeneous networks to represent such recommendation problems. We propose the weighted hybrid of low-dimensional recommenders (WHyLDR) recommendation model, which uses extended relations, represented as constrained network paths, to effectively augment direct relations. This model incorporates influences from both distant and proximal connections in the network. The WHyLDR approach raises the problem of the unconstrained proliferation of components, built from ever-extended network paths. We show that although component utility is not strictly monotonic with path length, a measure based on information gain can effectively prune and optimize such hybrids.

Recommender systems help users find items of interest by tailoring their recommendations to users' personal preferences. The utility of an item for a user, however, may vary greatly depending on that user's specific situation or the context in which the item is used. Without considering these changes in preferences, the recommendations may match the general preferences of a user, but they may have small value for the user in his/her current situation. In this paper, we introduce a hierarchical hidden Markov model for capturing changes in user's preferences. Using a user's feedback sequence on items, we model the user as a hierarchical hidden Markov process and the current context of the user as a hidden variable in this model. For a given user, our model is used to infer the maximum likelihood sequence of transitions between contextual states and to predict the probability distribution for the context of the next action. The predicted context is then used to generate recommendations. Our evaluation results using Last.fm music playlist data, indicate that this approach achieves significantly better performance in terms of accuracy and diversity compared to baseline methods.

Context-aware recommender systems (CARS) help improve the effectiveness of recommendations by adapting to users' preferences in different contextual situations. One approach to CARS that has been shown to be particularly effective is Context-Aware Matrix Factorization (CAMF). CAMF incorporates contextual dependencies into the standard matrix factorization (MF) process, where users and items are represented as collections of weights over various latent factors. In this paper, we introduce another CARS approach based on an extension of matrix factorization, namely, the Sparse Linear Method (SLIM). We develop a family of deviation-based contextual SLIM (CSLIM) recommendation algorithms by learning rating deviations in different contextual conditions. Our CSLIM approach is better at explaining the underlying reasons behind contextual recommendations, and our experimental evaluations over five context-aware data sets demonstrate that these CSLIM algorithms outperform the state-of-the-art CARS algorithms in the top-N recommendation task. We also discuss the criteria for selecting the appropriate CSLIM algorithm in advance based on the underlying characteristics of the data.

Contextual factors can greatly influence the utility of recommendations for users. In many recommendation and personalization applications, particularly in domains where user context changes dynamically, it is difficult to represent and model contextual factors directly, but it is often possible to observe their impact on user preferences during the course of users' interactions with the system. In this paper, we introduce an interactive recommender system that can detect and adapt to changes in context based on the user's ongoing behavior. The system, then, dynamically tailors its recommendations to match the user's most recent preferences. We formulate this problem as a multi-armed bandit problem and use Thompson sampling heuristic to learn a model for the user. Following the Thompson sampling approach, the user model is updated after each interaction as the system observes the corresponding rewards for the recommendations provided during that interaction. To generate contextual recommendations, the user's preference model is monitored for changes at each step of interaction with the user and is updated incrementally. We will introduce a mechanism for detecting significant changes in the user's preferences and will describe how it can be used to improve the performance of the recommender system.

Context-aware recommender systems (CARS) take contextual conditions into account when providing item recommendations. In recent years, context-aware matrix factorization (CAMF) has emerged as an extension of the matrix factorization technique that also incorporates contextual conditions. In this paper, we introduce another matrix factorization approach for contextual recommendations, the contextual SLIM (CSLIM) recommendation approach. It is derived from the sparse linear method (SLIM) which was designed for Top-N recommendations in traditional recommender systems. Based on the experimental evaluations over several context-aware data sets, we demonstrate that CLSIM can be an effective approach for context-aware recommendations, in many cases outperforming state-of-the-art CARS algorithms in the Top-N recommendation task.

Context aware recommender systems go beyond the traditional personalized recommendation models by incorporating a form of situational awareness. They provide recommendations that not only correspond to a user's preference profile, but that are also tailored to a given situation or context. We consider the setting in which contextual information is represented as a subset of an item feature space describing short-term interests or needs of a user in a given situation. This contextual information can be provided by the user in the form of an explicit query, or derived implicitly.

Group recommender systems (RS) are used to support groups in making common decisions when considering a set of alternatives. Current approaches generate group recommendations based on the users' individual preferences models. We believe that members of a group can reach an agreement more effectively by exchanging proposals suggested by a conventional RS. We propose to use a critiquing RS that has been shown to be effective in single-user recommendation. In the group recommendation context, critiquing allows each user to get new recommendations similar to the proposals made by the other group members and to communicate the rationale behind their own counterproposals. We describe a mobile application implementing the proposed approach and its evaluation in a live user experiment.

Social annotation systems allow users to annotate resources with personalized tags and to navigate large and complex information spaces without the need to rely on predefined hierarchies. These systems help users organize and share their own resources, as well as discover new ones annotated by other users. Tag recommenders in such systems assist users in finding appropriate tags for resources and help consolidate annotations across all users and resources. But the size and complexity of the data, as well as the inherent noise and inconsistencies in the underlying tag vocabularies, have made the design of effective tag recommenders a challenge. Recent efforts have demonstrated the advantages of integrative models that leverage all three dimensions of a social annotation system: users, resources and tags. Among these approaches are recommendation models based on matrix factorization. But, these models tend to lack scalability and often hide the underlying characteristics, or "information channels" of the data that affect recommendation effectiveness. In this paper we propose a weighted hybrid tag recommender that blends multiple recommendation components drawing separately on complementary dimensions, and evaluate it on six large real-world datasets. In addition, we attempt to quantify the strength of the information channels in these datasets and use these results to explain the performance of the hybrid. We find our approach is not only competitive with the state-of-the-art techniques in terms of accuracy, but also has the added benefits of being scalable to large real world applications, extensible to incorporate a wide range of recommendation techniques, easily updateable, and more scrutable than other leading methods.

Collaborative recommendation algorithms are typically evaluated on a static matrix of user rating data. However, when users experience a recommender system, it is dynamic, constantly evolving as new items and new users arrive. The dynamic properties of collaborative recommendation have become important as prediction algorithms based on the interactions of rating histories have been proposed, and as researchers seek to understand problems of robustness and maintenance in rating databases.

This tutorial will discuss vulnerabilities of collaborative recommendation algorithms: attacks that can be mounted against them and possible defenses that can be used. The tutorial will be of interest to researchers and practitioners in the area of collaborative recommendation.

Collaborative tagging applications allow Internet users to annotate resources with personalized tags. The complex network created by many annotations, often called a folksonomy, permits users the freedom to explore tags, resources or even other user's profiles unbound from a rigid predefined conceptual hierarchy. However, the freedom afforded users comes at a cost: an uncontrolled vocabulary can result in tag redundancy and ambiguity hindering navigation. Data mining techniques, such as clustering, provide a means to remedy these problems by identifying trends and reducing noise. Tag clusters can also be used as the basis for effective personalized recommendation assisting users in navigation. We present a personalization algorithm for recommendation in folksonomies which relies on hierarchical tag clusters. Our basic recommendation framework is independent of the clustering method, but we use a context-dependent variant of hierarchical agglomerative clustering which takes into account the user's current navigation context in cluster selection. We present extensive experimental results on two real world dataset. While the personalization algorithm is successful in both cases, our results suggest that folksonomies encompassing only one topic domain, rather than many topics, present an easier target for recommendation, perhaps because they are more focused and often less sparse. Furthermore, context dependent cluster selection, an integral step in our personalization algorithm, demonstrates more utility for recommendation in multi-topic folksonomies than in single-topic folksonomies. This observation suggests that topic selection is an important strategy for recommendation in multi-topic folksonomies.

Every user has a distinct background and a specific goal when searching for information on the Web. The goal of Web search personalization is to tailor search results to a particular user based on that user's interests and preferences. Effective personalization of information access involves two important challenges: accurately identifying the user context and organizing the information in such a way that matches the particular context. We present an approach to personalized search that involves building models of user context as ontological profiles by assigning implicitly derived interest scores to existing concepts in a domain ontology. A spreading activation algorithm is used to maintain the interest scores based on the user's ongoing behavior. Our experiments show that re-ranking the search results based on the interest scores and the semantic evidence in an ontological user profile is effective in presenting the most relevant results to the user.

Standard memory-based collaborative filtering algorithms, such as k-nearest neighbor, are quite vulnerable to profile injection attacks. Previous work has shown that some model-based techniques are more robust than k-nn. Model abstraction can inhibit certain aspects of an attack, providing an algorithmic approach to minimizing attack effectiveness. In this paper, we examine the robustness of a recommendation algorithm based on the data mining technique of association rule mining. Our results show that the Apriori algorithm offers large improvement in stability and robustness compared to k-nearest neighbor and other model-based techniques we have studied. Furthermore, our results show that Apriori can achieve comparable recommendation accuracy to k-nn.

Collaborative recommender systems are highly vulnerable to attack. Attackers can use automated means to inject a large number of biased profiles into such a system, resulting in recommendations that favor or disfavor given items. Since collaborative recommender systems must be open to user input, it is difficult to design a system that cannot be so attacked. Researchers studying robust recommendation have therefore begun to identify types of attacks and study mechanisms for recognizing and defeating them. In this paper, we propose and study different attributes derived from user profiles for their utility in attack detection. We show that a machine learning classification approach that includes attributes derived from attack models is more successful than more generalized detection algorithms previously studied.

Significant vulnerabilities have recently been identified in collaborative filtering recommender systems. Researchers have shown that attackers can manipulate a system's recommendations by injecting biased profiles into it. In this paper, we examine attacks that concentrate on a targeted set of users with similar tastes, biasing the system's responses to these users. We show that such attacks are both pragmatically reasonable and also highly effective against both user-based and item-based algorithms. As a result, an attacker can mount such a "segmented" attack with little knowledge of the specific system being targeted and with strong likelihood of success.