FairFS: Addressing Deep Feature Selection Biases for Recommender System

TL;DR

FairFS is a novel feature selection algorithm for recommender systems that mitigates biases in importance estimation, leading to more accurate and fair feature importance identification in deep learning models.

Contribution

This paper introduces FairFS, a new method that addresses layer, baseline, and approximation biases in deep feature selection for recommender systems.

Findings

FairFS outperforms existing methods in feature selection accuracy.

It effectively mitigates biases in importance estimation.

Achieves state-of-the-art performance in real-world experiments.

Abstract

Large-scale online marketplaces and recommender systems serve as critical technological support for e-commerce development. In industrial recommender systems, features play vital roles as they carry information for downstream models. Accurate feature importance estimation is critical because it helps identify the most useful feature subsets from thousands of feature candidates for online services. Such selection enables improved online performance while reducing computational cost. To address feature selection problems in deep learning, trainable gate-based and sensitivity-based methods have been proposed and proven effective in industrial practice. However, through the analysis of real-world cases, we identified three bias issues that cause feature importance estimation to rely on partial model layers, samples, or gradients, ultimately leading to inaccurate importance estimation. We refer to these as layer bias, baseline bias, and approximation bias. To mitigate these issues, we propose FairFS, a fair and accurate feature selection algorithm. FairFS regularizes feature importance estimated across all nonlinear transformation layers to address layer bias. It also introduces a smooth baseline feature close to the classifier decision boundary and adopts an aggregated approximation method to alleviate baseline and approximation biases. Extensive experiments demonstrate that FairFS effectively mitigates these biases and achieves state-of-the-art feature selection performance.