Fair Transfer Learning

Description

An in-processing fairness technique that adapts pre-trained models from one domain to another whilst explicitly preserving fairness constraints across different contexts or populations. The method addresses the challenge that fairness properties may not transfer when adapting models to new domains with different demographic compositions or data distributions. Fair transfer learning typically involves constraint-aware fine-tuning, domain adaptation techniques, or adversarial training that maintains equitable performance across groups in the target domain, ensuring that bias mitigation efforts carry over from source to target domains.

Example Use Cases

Fairness

Adapting a hiring algorithm trained on one country's recruitment data to another region whilst maintaining fairness across gender and ethnicity groups, ensuring equitable candidate evaluation despite different local demographic distributions and cultural contexts.

Transparency

Transferring a medical diagnosis model from urban hospital data to rural clinics whilst providing transparent evidence that fairness constraints are preserved across age, gender, and socioeconomic groups despite different patient populations and healthcare infrastructure.

Reliability

Adapting a fraud detection system from one financial market to another whilst ensuring reliable performance across customer demographics, maintaining consistent accuracy and fairness even when transaction patterns and customer characteristics differ between markets.

Limitations

Fairness properties achieved in the source domain may not translate directly to the target domain if demographic distributions or data characteristics differ significantly.
Requires careful hyperparameter tuning and constraint specification to balance fairness preservation with model performance in the new domain.
Implementation complexity is high, requiring expertise in both transfer learning techniques and fairness constraint optimisation methods.
May suffer from negative transfer effects where fairness constraints that worked well in the source domain actually harm performance in the target domain.
Evaluation challenges arise from needing to validate fairness across multiple domains and demographic groups simultaneously.

Resources

Research Papers

Segmenting across places: The need for fair transfer learning with satellite imagery

Miao Zhang et al.•Apr 9, 2022

The increasing availability of high-resolution satellite imagery has enabled the use of machine learning to support land-cover measurement and inform policy-making. However, labelling satellite images is expensive and is available for only some locations. This prompts the use of transfer learning to adapt models from data-rich locations to others. Given the potential for high-impact applications of satellite imagery across geographies, a systematic assessment of transfer learning implications is warranted. In this work, we consider the task of land-cover segmentation and study the fairness implications of transferring models across locations. We leverage a large satellite image segmentation benchmark with 5987 images from 18 districts (9 urban and 9 rural). Via fairness metrics we quantify disparities in model performance along two axes -- across urban-rural locations and across land-cover classes. Findings show that state-of-the-art models have better overall accuracy in rural areas compared to urban areas, through unsupervised domain adaptation methods transfer learning better to urban versus rural areas and enlarge fairness gaps. In analysis of reasons for these findings, we show that raw satellite images are overall more dissimilar between source and target districts for rural than for urban locations. This work highlights the need to conduct fairness analysis for satellite imagery segmentation models and motivates the development of methods for fair transfer learning in order not to introduce disparities between places, particularly urban and rural locations.

Trustworthy Transfer Learning: A Survey

Jun Wu and Jingrui He•Jan 1, 2024

Transfer learning aims to transfer knowledge or information from a source domain to a relevant target domain. In this paper, we understand transfer learning from the perspectives of knowledge transferability and trustworthiness. This involves two research questions: How is knowledge transferability quantitatively measured and enhanced across domains? Can we trust the transferred knowledge in the transfer learning process? To answer these questions, this paper provides a comprehensive review of trustworthy transfer learning from various aspects, including problem definitions, theoretical analysis, empirical algorithms, and real-world applications. Specifically, we summarize recent theories and algorithms for understanding knowledge transferability under (within-domain) IID and non-IID assumptions. In addition to knowledge transferability, we review the impact of trustworthiness on transfer learning, e.g., whether the transferred knowledge is adversarially robust or algorithmically fair, how to transfer the knowledge under privacy-preserving constraints, etc. Beyond discussing the current advancements, we highlight the open questions and future directions for understanding transfer learning in a reliable and trustworthy manner.

Cross-Institutional Transfer Learning for Educational Models: Implications for Model Performance, Fairness, and Equity

Josh Gardner et al.•May 1, 2023

Modern machine learning increasingly supports paradigms that are multi-institutional (using data from multiple institutions during training) or cross-institutional (using models from multiple institutions for inference), but the empirical effects of these paradigms are not well understood. This study investigates cross-institutional learning via an empirical case study in higher education. We propose a framework and metrics for assessing the utility and fairness of student dropout prediction models that are transferred across institutions. We examine the feasibility of cross-institutional transfer under real-world data- and model-sharing constraints, quantifying model biases for intersectional student identities, characterizing potential disparate impact due to these biases, and investigating the impact of various cross-institutional ensembling approaches on fairness and overall model performance. We perform this analysis on data representing over 200,000 enrolled students annually from four universities without sharing training data between institutions. We find that a simple zero-shot cross-institutional transfer procedure can achieve similar performance to locally-trained models for all institutions in our study, without sacrificing model fairness. We also find that stacked ensembling provides no additional benefits to overall performance or fairness compared to either a local model or the zero-shot transfer procedure we tested. We find no evidence of a fairness-accuracy tradeoff across dozens of models and transfer schemes evaluated. Our auditing procedure also highlights the importance of intersectional fairness analysis, revealing performance disparities at the intersection of sensitive identity groups that are concealed under one-dimensional analysis.

Related Techniques

Name	Description	Assurance Goals
Adversarial Debiasing	Adversarial debiasing reduces bias by training models using a competitive adversarial setup, similar to Generative Adversarial Networks (GANs). The technique involves two neural networks: a predictor that learns to make accurate predictions on the main task, and an adversary (bias detector) that attempts to predict protected attributes (such as race, gender, or age) from the predictor's internal representations. Through adversarial training, the predictor learns to produce representations that retain predictive power for the main task whilst being uninformative about protected characteristics, thereby reducing discriminatory bias.	Fairness
Fair Adversarial Networks	An in-processing fairness technique that employs adversarial training with dual neural networks to learn fair representations. The method consists of a predictor network that learns the main task whilst an adversarial discriminator network simultaneously attempts to predict sensitive attributes from the predictor's hidden representations. Through this adversarial min-max game, the predictor is incentivised to learn features that are informative for the task but statistically independent of protected attributes, effectively removing bias at the representation level in deep learning models.	Fairness Transparency Reliability
Exponentiated Gradient Reduction	An in-processing fairness technique based on Agarwal et al.'s reductions approach that transforms fair classification into a sequence of cost-sensitive classification problems. The method uses an exponentiated gradient algorithm to iteratively reweight training data, returning a randomised classifier that achieves the lowest empirical error whilst satisfying fairness constraints. This reduction-based framework provides theoretical guarantees about both accuracy and constraint violation, making it suitable for various fairness criteria including demographic parity and equalised odds.	Fairness Transparency Reliability
Adaptive Sensitive Reweighting	Adaptive Sensitive Reweighting dynamically adjusts the importance of training examples during model training based on real-time performance across different demographic groups. Unlike traditional static reweighting that fixes weights at the start, this technique continuously monitors fairness metrics and automatically increases the weight of examples from underperforming groups whilst decreasing weights for overrepresented groups. The adaptive mechanism prevents models from perpetuating historical biases by ensuring balanced learning across all demographics throughout the training process.	Fairness Reliability