FairBatch: Batch Selection for Model Fairness

By Yuji Roh, Kangwook Lee, Steven Euijong Whang, Changho Suh. 2020

163 citations recorded by the source corpus.

Training a fair machine learning model is essential to prevent demographic disparity. Existing techniques for improving model fairness require broad changes in either data preprocessing or model training, rendering themselves difficult-to-adopt for potentially already complex machine learning systems. We address this problem via the lens of bilevel optimization. While keeping the standard training algorithm as an inner optimizer, we incorporate an outer optimizer so as to equip the inner problem with an additional functionality: Adaptively selecting minibatch sizes for the purpose of improving model fairness. Our batch selection algorithm, which we call FairBatch, implements this optimization and supports prominent fairness measures: equal opportunity, equalized odds, and demographic parity. FairBatch comes with a significant implementation benefit -- it does not require any modification to data preprocessing or model training. For instance, a single-line change of PyTorch code for replacing batch selection part of model training suffices to employ FairBatch. Our experiments conducted both on synthetic and benchmark real data demonstrate that FairBatch can provide such functionalities while achieving comparable (or even greater) performances against the state of the arts. Furthermore, FairBatch can readily improve fairness of any pre-trained model simply via fine-tuning. It is also compatible with existing batch selection techniques intended for different purposes, such as faster convergence, thus gracefully achieving multiple purposes.

Subfields: Artificial Intelligence, Safety Research, Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Safety Research

In Safety, Robustness & Interpretability

Citations in the corpus

On The Impact of Machine Learning Randomness on Group Fairness (2023, background): “Fairbatch [46], a recently proposed bias mitigation algorithm, follows a similar formulation (although it additionally changes the overall distribution).”
Debiasing Pre-Trained Language Models via Efficient Fine-Tuning (2022, methodology): “To mitigate biased behaviors in machine learning models, researchers have proposed meth-ods targeting different tasks and domains, such as classification (Menon and Williamson, 2018; Roh et al., 2021), regression (Agarwal et al., 2019; Berk et al., 2017), and adversarial learning (Xu et al., 2018).”
Fairness in Machine Learning: A Survey (2020, background): “In gradient descent-based models, several in-processing approaches have been suggested to learn instance weights such that fairness can be improved [149, 207, 227, 246].”
LIFT: Language-Interfaced Fine-Tuning for Non-Language Machine Learning Tasks (2022, methodology): “To further remove the bias, users can combine LIFT with the existing fairness-aware reweighting mechanisms [179, 180] or data augmentation and parameter-efficient fine-tuning techniques [181].”
FairFed: Enabling Group Fairness in Federated Learning (2021, unknown): “In (Zeng, Chen, and Lee 2021), an adaptation of the Fair-Batch debiasing algorithm (Roh et al. 2021) is proposed for FL where clients use FairBatch locally and the weights are updated through the server in each round.”
Data collection and quality challenges in deep learning: a data-centric AI perspective (2021, methodology): “Instead, the idea of FairBatch is to adjust the sensitive group ratios within each batch of examples being used for training as illustrated in Figure 16.”