A Comparative Analysis of Bias Amplification in Graph Neural Network Approaches for Recommender Systems

Chizari, Nikzad; Shoeibi, Niloufar; García, María N. Moreno

doi:10.3390/electronics11203301

Cited by 14 publications

(8 citation statements)

References 91 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is claimed that the use of the GNN-based method for RS can improve the accuracy of the results [18][19][20]. On the other hand, this enhancement in performance can cause bias and fairness problems [21,22]. As discussed in Section 2.2, the structure of graphs together with the message-passing system inside GNNs can amplify bias problems, leading to unfair outcomes.…”

Section: Bias and Fairness In Gnn-based Rsmentioning

confidence: 99%

“…Within this approach, the methods based on graph neural networks (GNNs) perform well in this application domain, although they are still affected by the bias shortcoming. Some studies have shown that graph structures and the used message-passing system inside GNNs promote the amplification of unfairness and other social biases [21,22]. Moreover, in most of the social networks with graph architecture, nodes with similar sensitive attributes are prone to be connected in comparison to other nodes with different sensitive attributes (e.g., young individuals are more likely to start a friendship in social networks).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Bias Assessment Approaches for Addressing User-Centered Fairness in GNN-Based Recommender Systems

2023

Self Cite

View full text Add to dashboard Cite

In today’s technology-driven society, many decisions are made based on the results provided by machine learning algorithms. It is widely known that the models generated by such algorithms may present biases that lead to unfair decisions for some segments of the population, such as minority or marginalized groups. Hence, there is concern about the detection and mitigation of these biases, which may increase the discriminatory treatments of some demographic groups. Recommender systems, used today by millions of users, are not exempt from this drawback. The influence of these systems on so many user decisions, which in turn are taken as the basis for future recommendations, contributes to exacerbating this problem. Furthermore, there is evidence that some of the most recent and successful recommendation methods, such as those based on graphical neural networks (GNNs), are more sensitive to bias. The evaluation approaches of some of these biases, as those involving protected demographic groups, may not be suitable for recommender systems since their results are the preferences of the users and these do not necessarily have to be the same for the different groups. Other assessment metrics are aimed at evaluating biases that have no impact on the user. In this work, the suitability of different user-centered bias metrics in the context of GNN-based recommender systems are analyzed, as well as the response of recommendation methods with respect to the different types of biases to which these measures are addressed.

show abstract

Section: Bias and Fairness In Gnn-based Rsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Bias Assessment Approaches for Addressing User-Centered Fairness in GNN-Based Recommender Systems

2023

Self Cite

View full text Add to dashboard Cite

show abstract

“…Fairness helps to mitigate bias, supports diversity, and boosts user satisfaction. In GNN-based systems, which can amplify bias, fairness is crucial for balanced recommendations and optimal performance (Ekstrand et al, 2018 ; Chizari et al, 2022 ; Chen et al, 2023 ; Gao et al, 2023 ).…”

Section: Fairness In Gnn-based Recommender Systemsmentioning

confidence: 99%

Beyond-accuracy: a review on diversity, serendipity, and fairness in recommender systems based on graph neural networks

Duricic,

Kowald,

Lacic

et al. 2023

Front. Big Data

View full text Add to dashboard Cite

By providing personalized suggestions to users, recommender systems have become essential to numerous online platforms. Collaborative filtering, particularly graph-based approaches using Graph Neural Networks (GNNs), have demonstrated great results in terms of recommendation accuracy. However, accuracy may not always be the most important criterion for evaluating recommender systems' performance, since beyond-accuracy aspects such as recommendation diversity, serendipity, and fairness can strongly influence user engagement and satisfaction. This review paper focuses on addressing these dimensions in GNN-based recommender systems, going beyond the conventional accuracy-centric perspective. We begin by reviewing recent developments in approaches that improve not only the accuracy-diversity trade-off but also promote serendipity, and fairness in GNN-based recommender systems. We discuss different stages of model development including data preprocessing, graph construction, embedding initialization, propagation layers, embedding fusion, score computation, and training methodologies. Furthermore, we present a look into the practical difficulties encountered in assuring diversity, serendipity, and fairness, while retaining high accuracy. Finally, we discuss potential future research directions for developing more robust GNN-based recommender systems that go beyond the unidimensional perspective of focusing solely on accuracy. This review aims to provide researchers and practitioners with an in-depth understanding of the multifaceted issues that arise when designing GNN-based recommender systems, setting our work apart by offering a comprehensive exploration of beyond-accuracy dimensions.

show abstract

“…As mentioned in the introduction, the theory of the spiral of silence is of significant importance in the climate of recommender systems, as the standard for recommendations in different recommendation systems can vary significantly due to various user attributes, such as race, gender, and personality [4].…”

Section: Survey Of Relevant Previous Workmentioning

confidence: 99%

Effects of the spiral of silence on minority groups in recommender systems

Khan,

Herder

2023

Proceedings of the 34th ACM Conference on Hypertext and Social Media

View full text Add to dashboard Cite

Recommender systems play a critical role in today's data-rich landscape, where the abundance of information necessitates their ability to present the most relevant choices for individual users. However, it has been noted that recommender systems often fail to offer the most suitable options for minority groups. Therefore, this paper examines one of the potential reasons behind the underrepresentation of minority opinions within recommender systems through a literature review. Specifically, we explore the connection and importance between the spiral of silence concept and the dearth of suitable recommendations for minorities. CCS CONCEPTS• Human-centered computing → Interaction paradigms; Empirical studies in HCI .

show abstract

A Comparative Analysis of Bias Amplification in Graph Neural Network Approaches for Recommender Systems

Cited by 14 publications

References 91 publications

Bias Assessment Approaches for Addressing User-Centered Fairness in GNN-Based Recommender Systems

Bias Assessment Approaches for Addressing User-Centered Fairness in GNN-Based Recommender Systems

Beyond-accuracy: a review on diversity, serendipity, and fairness in recommender systems based on graph neural networks

Effects of the spiral of silence on minority groups in recommender systems

Contact Info

Product

Resources

About