# Feature-Wise Bias Amplification

arXiv: https://arxiv.org/abs/1812.08999 · HTML: https://mlsanity.com/paper/1812.08999/

## Key facts

- Cited by 31 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 21 Dec 2018.

## Built on by (most cited first, 30 of 31)

- [Gender bias and stereotypes in Large Language Models](https://mlsanity.com/paper/2308.14921/) (2023)
- [A Systematic Study of Bias Amplification](https://mlsanity.com/paper/2201.11706/) (2022)
- [Data Feedback Loops: Model-driven Amplification of Dataset Biases](https://mlsanity.com/paper/2209.03942/) (2022)
- [Directional Bias Amplification](https://mlsanity.com/paper/2102.12594/) (2021)
- [Interactive Natural Language Processing](https://mlsanity.com/paper/2305.13246/) (2023)
- [Influence Tuning: Demoting Spurious Correlations via Instance Attribution and Instance-Driven Updates](https://mlsanity.com/paper/2110.03212/) (2021)
- [Men Also Do Laundry: Multi-Attribute Bias Amplification](https://mlsanity.com/paper/2210.11924/) (2022)
- [Evaluating Biased Attitude Associations of Language Models in an Intersectional Context](https://mlsanity.com/paper/2307.03360/) (2023)
- [Leashing the Inner Demons: Self-Detoxification for Language Models](https://mlsanity.com/paper/2203.03072/) (2022)
- [MedEval: A Multi-Level, Multi-Task, and Multi-Domain Medical Benchmark for Language Model Evaluation](https://mlsanity.com/paper/2310.14088/) (2023)
- [Model-Agnostic Gender Debiased Image Captioning](https://mlsanity.com/paper/2304.03693/) (2023)
- [Causal Intervention Improves Implicit Sentiment Analysis](https://mlsanity.com/paper/2208.09329/) (2022)
- [Co2PT: Mitigating Bias in Pre-trained Language Models through Counterfactual Contrastive Prompt Tuning](https://mlsanity.com/paper/2310.12490/) (2023)
- [Harm Amplification in Text-to-Image Models](https://mlsanity.com/paper/2402.01787/) (2024)
- [Escaping the SpuriVerse: Can Large Vision-Language Models Generalize Beyond Seen Spurious Correlations?](https://mlsanity.com/paper/2506.18322/) (2025)
- [Simplicity Bias Leads to Amplified Performance Disparities](https://mlsanity.com/paper/2212.06641/) (2022)
- [Automating Data Annotation under Strategic Human Agents: Risks and Potential Solutions](https://mlsanity.com/paper/2405.08027/) (2024)
- [Towards an Enhanced Understanding of Bias in Pre-trained Neural Language Models: A Survey with Special Emphasis on Affective Bias](https://mlsanity.com/paper/2204.10365/) (2022)
- [Synthetic Pre-Training Tasks for Neural Machine Translation](https://mlsanity.com/paper/2212.09864/) (2022)
- [Toward Operationalizing Pipeline-aware ML Fairness: A Research Agenda for Developing Practical Guidelines and Tools](https://mlsanity.com/paper/2309.17337/) (2023)
- [CAT: Controllable Attribute Translation for Fair Facial Attribute Classification](https://mlsanity.com/paper/2209.06850/) (2022)
- [Addressing Discretization-Induced Bias in Demographic Prediction](https://mlsanity.com/paper/2405.16762/) (2024)
- [Protected group bias and stereotypes in Large Language Models](https://mlsanity.com/paper/2403.14727/) (2024)
- [Gradient Based Activations for Accurate Bias-Free Learning](https://mlsanity.com/paper/2202.10943/) (2022)
- [Neural Conditional Transport Maps](https://mlsanity.com/paper/2505.15808/) (2025)
- [A Modulation Layer to Increase Neural Network Robustness Against Data Quality Issues](https://mlsanity.com/paper/2107.08574/) (2021)
- [Towards Fair and Calibrated Models](https://mlsanity.com/paper/2310.10399/) (2023)
- [In-Context Bias Propagation in LLM-Based Tabular Data Generation](https://mlsanity.com/paper/2506.09630/) (2025)
- [Validity of Feature Importance in Low-Performing Machine Learning for Tabular Biomedical Data](https://mlsanity.com/paper/2409.13342/) (2024)
- [A Critical Review of Predominant Bias in Neural Networks](https://mlsanity.com/paper/2502.11031/) (2025)

## Abstract (arXiv)

We study the phenomenon of bias amplification in classifiers, wherein a machine learning model learns to predict classes with a greater disparity than the underlying ground truth. We demonstrate that bias amplification can arise via an inductive bias in gradient descent methods that results in the overestimation of the importance of moderately-predictive "weak" features if insufficient training data is available. This overestimation gives rise to feature-wise bias amplification -- a previously unreported form of bias that can be traced back to the features of a trained model. Through analysis and experiments, we show that while some bias cannot be mitigated without sacrificing accuracy, feature-wise bias amplification can be mitigated through targeted feature selection. We present two new feature selection algorithms for mitigating bias amplification in linear models, and show how they can be adapted to convolutional neural networks efficiently. Our experiments on synthetic and real data demonstrate that these algorithms consistently lead to reduced bias without harming accuracy, in some cases eliminating predictive bias altogether while providing modest gains in accuracy.