# FairFS: Addressing Deep Feature Selection Biases for Recommender System

arXiv: https://arxiv.org/abs/2602.20001 · HTML: https://mlsanity.com/paper/2602.20001/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 61 references that are in the map.
- First posted on arXiv 23 Feb 2026.

## Builds on (most cited first, 24 of 24)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Categorical Reparameterization with Gumbel-Softmax](https://mlsanity.com/paper/1611.01144/) (2016)
- [SNIP: Single-shot Network Pruning based on Connection Sensitivity](https://mlsanity.com/paper/1810.02340/) (2018)
- [Wide & Deep Learning for Recommender Systems](https://mlsanity.com/paper/1606.07792/) (2016)
- [DeepFM: A Factorization-Machine based Neural Network for CTR Prediction](https://mlsanity.com/paper/1703.04247/) (2017)
- [Interpretable Explanations of Black Boxes by Meaningful Perturbation](https://mlsanity.com/paper/1704.03296/) (2017)
- [Deep & Cross Network for Ad Click Predictions](https://mlsanity.com/paper/1708.05123/) (2017)
- [AutoInt: Automatic Feature Interaction Learning via Self-Attentive Neural Networks](https://mlsanity.com/paper/1810.11921/) (2018)
- [xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems](https://mlsanity.com/paper/1803.05170/) (2018)
- [Product-based Neural Networks for User Response Prediction](https://mlsanity.com/paper/1611.00144/) (2016)
- [LassoNet: A Neural Network with Feature Sparsity](https://mlsanity.com/paper/1907.12207/) (2019)
- [OpenFE: Automated Feature Generation with Expert-level Performance](https://mlsanity.com/paper/2211.12507/) (2022)
- [COLD: Towards the Next Generation of Pre-Ranking System](https://mlsanity.com/paper/2007.16122/) (2020)
- [AutoField: Automating Feature Selection in Deep Recommender Systems](https://mlsanity.com/paper/2204.09078/) (2022)
- [Towards a Better Tradeoff between Effectiveness and Efficiency in Pre-Ranking: A Learnable Feature Selection based Approach](https://mlsanity.com/paper/2105.07706/) (2021)
- [Multi-Interactive Attention Network for Fine-grained Feature Learning in CTR Prediction](https://mlsanity.com/paper/2012.06968/) (2020)
- [Dropout Feature Ranking for Deep Learning Models](https://mlsanity.com/paper/1712.08645/) (2017)
- [MadaKV: Adaptive Modality-Perception KV Cache Eviction for Efficient Multimodal Long-Context Inference](https://mlsanity.com/paper/2506.15724/) (2025)
- [MvFS: Multi-view Feature Selection for Recommender System](https://mlsanity.com/paper/2309.02064/) (2023)
- [SHARK: A Lightweight Model Compression Approach for Large-scale Recommender Systems](https://mlsanity.com/paper/2308.09395/) (2023)
- [MergeNet: Knowledge Migration across Heterogeneous Models, Tasks, and Modalities](https://mlsanity.com/paper/2404.13322/) (2024)
- [SELF: Surrogate-light Feature Selection with Large Language Models in Deep Recommender Systems](https://mlsanity.com/paper/2412.08516/) (2024)
- [Revisiting Feature Interactions from the Perspective of Quadratic Neural Networks for Click-through Rate Prediction](https://mlsanity.com/paper/2505.17999/) (2025)
- [TayFCS: Towards Light Feature Combination Selection for Deep Recommender Systems](https://mlsanity.com/paper/2507.03895/) (2025)

## Abstract (arXiv)

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.