Learning with Exposure Constraints in Recommendation Systems

Ben-Porat, Omer; Rotem, Torkan,

doi:10.1145/3543507.3583320

Cited by 7 publications

(3 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Yang and Ai [40] proposed a marginal optimizing approach to conduct amortized MMF in the learning-to-rank process. TFROM [37] and CP-Fair [26] proposed a Linear Programming (LP)-based method to ensure the group fairness, see also [2,9,41]. P-MMF [38], LTP-MMF [39] proposed an online mirror gradient descent to improve worst-off provider's exposures in the dual space.…”

Section: Related Workmentioning

confidence: 99%

“…In the original space (ComiRec-DR embeddings), it is evident that the user embeddings are closely aligned with the embeddings of category 3. However, in the dual space (ComiRec-DR+FairSync embeddings), the user embeddings are in closer proximity to lower categories (1,2), thereby ensuring that other categories meet the minimum exposure requirements.…”

Section: Visualization Of Embeddings Under Original and Dual Spacementioning

confidence: 99%

“…In recent times, the need for fair recommender systems (RS) has gained prominence in industrial requirements [21,33]. Among these requirements, RS platform has the demands of guaranteeing that specific groups achieve a minimum level of exposure to items within a defined period, aligning with the perspective of amortized group max-min fairness (MMF) [2,18,25,38,41]. For instance, certain studies propose to ensure minimum item exposures for new providers [4,28,38,39] for attracting providers to join, while others focus on enhancing the visibility of specific item categories [45] for promoting certain festivals.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

FairSync: Ensuring Amortized Group Exposure in Distributed Recommendation Retrieval

Xu,

Ding

et al. 2024

Proceedings of the ACM Web Conference 2024

View full text Add to dashboard Cite

In pursuit of fairness and balanced development, recommender systems (RS) often prioritize group fairness, ensuring that specific groups maintain a minimum level of exposure over a given period. For example, RS platforms aim to ensure adequate exposure for new providers or specific categories of items according to their needs. Modern industry RS usually adopts a two-stage pipeline: stage-1 (retrieval stage) retrieves hundreds of candidates from millions of items distributed across various servers, and stage-2 (ranking stage) focuses on presenting a small-size but accurate selection from items chosen in stage-1. Existing efforts for ensuring amortized group exposures focus on stage-2, however, stage-1 is also critical for the task. Without a high-quality set of candidates, the stage-2 ranker cannot ensure the required exposure of groups. Previous fairnessaware works designed for stage-2 typically require accessing and traversing all items. In stage-1, however, millions of items are distributively stored in servers, making it infeasible to traverse all of them. How to ensure group exposures in the distributed retrieval process is a challenging question. To address this issue, we introduce a model named FairSync, which transforms the problem into a constrained distributed optimization problem. Specifically, FairSync resolves the issue by moving it to the dual space, where a central node aggregates historical fairness data into a vector and distributes it to all servers. To trade off the efficiency and accuracy, the gradient descent technique is used to periodically update the parameter of the dual vector. The experiment results on two public recommender retrieval datasets showcased that FairSync outperformed all the baselines, achieving the desired minimum level of exposures while maintaining a high level of retrieval accuracy.

show abstract