Federated learning (FL) provides a privacy-preserving approach for collaborative training of machine learning models. Given the potential data heterogeneity, it is crucial to select appropriate collaborators for each FL participant (FL-PT) based on data complementarity. Recent studies have addressed this challenge. Similarly, it is imperative to consider the inter-individual relationships among FL-PTs where some FL-PTs engage in competition. Although FL literature has acknowledged the significance of this scenario, practical methods for establishing FL ecosystems remain largely unexplored. In this paper, we extend a principle from the balance theory, namely ``the friend of my enemy is my enemy'', to ensure the absence of conflicting interests within an FL ecosystem. The extended principle and the resulting problem are formulated via graph theory and integer linear programming. A polynomial-time algorithm is proposed to determine the collaborators of each FL-PT. The solution guarantees high scalability, allowing even competing FL-PTs to smoothly join the ecosystem without conflict of interest. The proposed framework jointly considers competition and data heterogeneity. Extensive experiments on real-world and synthetic data demonstrate its efficacy compared to five alternative approaches, and its ability to establish efficient collaboration networks among FL-PTs.
Federated learning (FL) enables multiple participants (PTs) to build an aggregate and more powerful learning model without sharing data, thus maintaining data privacy and security. Among the key application scenarios is a competitive market where market shares represent PTs' competitiveness. An understanding of the role of FL in evolving market shares plays a key role in advancing the adoption of FL by PTs. In terms of modeling, we adapt a general economic model to the FL context and introduce two notions of $\delta$-stable market and friendliness to measure the viability of FL and the market acceptability to FL. Further, we address related decision-making issues with FL designer and PTs. First, we characterize the process by which each PT participates in FL as a non-cooperative game and prove its dominant strategy. Second, as an FL designer, the final model performance improvement of each PT should be bounded, which relates to the market conditions of a particular FL application scenario; we give a sufficient and necessary condition $Q$ to maintain the market $\delta$-stability and quantify the friendliness $\kappa$. The condition $Q$ gives a specific requirement while an FL designer allocates performance improvements among PTs. In a typical case of oligopoly, closed-form expressions of $Q$ and $\kappa$ are given. Finally, numerical results are given to show the viability of FL in a wide range of market conditions. Our results help identify optimal PT strategies, the viable operational space of an FL designer, and the market conditions under which FL is especially beneficial.