A Survey on Contribution Evaluation in Vertical Federated Learning
Yue Cui, Chung-ju Huang, Yuzhu Zhang, Leye Wang, Lixin Fan, Xiaofang Zhou, Qiang Yang
TL;DR
This survey addresses the challenge of fairly and accurately evaluating each participant’s contribution in vertical federated learning (VFL). It introduces a comprehensive taxonomy spanning the VFL lifecycle, granularity, privacy considerations, core evaluation methods, and application-specific tasks, with Shapley-value, leave-one-out, individual-based, and interaction-based approaches as the primary categories. The authors discuss data preprocessing, model training, and model inference as distinct CE phases, analyze privacy protocols (P-0 to P-3), and examine CE applications in feature selection, interpretable VFL, incentive design, and payment allocation; they also identify open challenges in privacy, communication, fairness, multi-objective optimization, and data valuation. The work provides practical guidance for designing privacy-centric, efficient, and fair contribution evaluation systems and points to open research directions and a living GitHub resource for ongoing updates. By organizing existing literature and proposing a unified framework, the paper aims to guide researchers and practitioners toward more equitable and sustainable VFL solutions.
Abstract
Vertical Federated Learning (VFL) has emerged as a critical approach in machine learning to address privacy concerns associated with centralized data storage and processing. VFL facilitates collaboration among multiple entities with distinct feature sets on the same user population, enabling the joint training of predictive models without direct data sharing. A key aspect of VFL is the fair and accurate evaluation of each entity's contribution to the learning process. This is crucial for maintaining trust among participating entities, ensuring equitable resource sharing, and fostering a sustainable collaboration framework. This paper provides a thorough review of contribution evaluation in VFL. We categorize the vast array of contribution evaluation techniques along the VFL lifecycle, granularity of evaluation, privacy considerations, and core computational methods. We also explore various tasks in VFL that involving contribution evaluation and analyze their required evaluation properties and relation to the VFL lifecycle phases. Finally, we present a vision for the future challenges of contribution evaluation in VFL. By providing a structured analysis of the current landscape and potential advancements, this paper aims to guide researchers and practitioners in the design and implementation of more effective, efficient, and privacy-centric VFL solutions. Relevant literature and open-source resources have been compiled and are being continuously updated at the GitHub repository: \url{https://github.com/cuiyuebing/VFL_CE}.