DICE: Data Influence Cascade in Decentralized Learning

2025-07-09

AI Summary

In decentralized learning, quantifying data influence across sparsely connected nodes remains challenging due to localized communication topologies, impeding fair contribution assessment and incentive design. This paper introduces the first cascade modeling framework for data influence in decentralized networks, theoretically establishing that influence propagation is jointly governed by data distribution, communication topology, and loss function curvature. Building on this analysis, we propose a computationally tractable multi-hop influence approximation method enabling node-level contribution attribution. The resulting traceable influence framework provides principled foundations for partner selection and adversarial behavior detection, while also enabling verifiable, incentive-compatible mechanism design. Empirical evaluation demonstrates significant improvements in collaboration incentives and system robustness.

Technology Category

• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Search and Optimization: Distributed Search
• Machine Learning: Causal Learning

Application Category

• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks

Abstract

Decentralized learning offers a promising approach to crowdsource data consumptions and computational workloads across geographically distributed compute interconnected through peer-to-peer networks, accommodating the exponentially increasing demands. However, proper incentives are still in absence, considerably discouraging participation. Our vision is that a fair incentive mechanism relies on fair attribution of contributions to participating nodes, which faces non-trivial challenges arising from the localized connections making influence ``cascade'' in a decentralized network. To overcome this, we design the first method to estimate extbf{D}ata extbf{I}nfluence extbf{C}ascad extbf{E} (DICE) in a decentralized environment. Theoretically, the framework derives tractable approximations of influence cascade over arbitrary neighbor hops, suggesting the influence cascade is determined by an interplay of data, communication topology, and the curvature of loss landscape. DICE also lays the foundations for applications including selecting suitable collaborators and identifying malicious behaviors. Project page is available at https://raiden-zhu.github.io/blog/2025/DICE/.

Problem

• Estimating data influence cascade in decentralized learning networks
• Designing fair incentive mechanisms for decentralized participation
• Addressing challenges from localized connections and influence cascades

Innovation

• Estimates data influence in decentralized networks
• Uses tractable approximations for influence cascade
• Determines influence by data, topology, loss curvature