In this essay, I explore ethical considerations that might arise from the use of collaborative filtering algorithms on dating apps. Collaborative filtering algorithms can predict the preferences of a target user by looking at the past behavior of similar users. By recommending products through this process, they can influence the news we read, the movies we watch, and more. They are extremely powerful and effective on platforms like Amazon and Google. Recommender systems on dating apps are likely to group people by race, since they exhibit similar patterns of behavior: users on dating platforms seem to segregate themselves based on race, exclude certain races from romantic and sexual consideration (except their own), and generally show a preference for white men and women. As collaborative filtering algorithms learn from these patterns to predict preferences and build recommendations, they can homogenize the behavior of dating app users and exacerbate biased sexual and romantic behavior.

Standard techniques in matrix factorization (MF) – a popular method for latent factor model-based design – result in dense matrices for both users and items. Users are likely to have some affinity toward all the latent factors – making a dense matrix plausible, but it is not possible for the items to possess all the latent factors simultaneously; hence it is more likely to be sparse. Therefore, we propose to factor the rating matrix into a dense user matrix and a sparse item matrix, leading to the blind compressed sensing (BCS) framework. To further enhance the prediction quality of our design, we aim to incorporate user and item metadata into the BCS framework. The additional information helps in reducing the underdetermined nature of the problem of rating prediction caused by extreme sparsity of the rating dataset. Our design is based on the belief that users sharing similar demographic profile have similar preferences and thus can be described by the similar latent factor vectors. We also use item metadata (genre information) to group together the similar items. We modify our BCS formulation to include item metadata under the assumption that items belonging to common genre share similar sparsity pattern. We also design an efficient algorithm to solve our formulation. Extensive experimentation conducted on the movielens dataset validates our claim that our modified MF framework utilizing auxiliary information improves upon the existing state-of-the-art techniques.