Collaboration Equilibrium in Federated Learning

Cui, Sen; Liang, Jian; Pan, Weishen; Chen, Kun; Zhang, Changshui; Wang, Fei

doi:10.1145/3534678.3539237

Cited by 11 publications

(15 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These incentives can include reputation, monetary compensation, or additional computational infrastructure, among others. 146 , 147…”

Section: Challengesmentioning

confidence: 99%

“…propose the concept of collaboration equilibrium, where clients are grouped such that no individual client could gain more in another configuration. 146 They employ a Pareto optimization framework and benefit graphs to create clusters of clients that reach this equilibrium. Although this approach exhibits potential for achieving collaborative fairness, it necessitates all local clients’ consent to construct a benefit graph by a neutral third party before the initiation of model training.…”

Section: Challengesmentioning

confidence: 99%

“…Although this approach exhibits potential for achieving collaborative fairness, it necessitates all local clients’ consent to construct a benefit graph by a neutral third party before the initiation of model training. 146 …”

Section: Challengesmentioning

confidence: 99%

See 2 more Smart Citations

Learning across diverse biomedical data modalities and cohorts: Challenges and opportunities for innovation

Rajendran,

Pan,

Sabuncu

et al. 2024

Patterns

View full text Add to dashboard Cite

“…These incentives can include reputation, monetary compensation, or additional computational infrastructure, among others. 146 , 147…”

Section: Challengesmentioning

confidence: 99%

Section: Challengesmentioning

confidence: 99%

See 1 more Smart Citation

Learning across diverse biomedical data modalities and cohorts: Challenges and opportunities for innovation

Rajendran,

Pan,

Sabuncu

et al. 2024

Patterns

View full text Add to dashboard Cite

“…For model fairness, Hu et al [ 33 ] proposed the FedMGDA+ method, which performs multi-objective optimization by optimizing the loss function of each FL client individually and simultaneously to avoid sacrificing the model performance of any client. Cui et al [ 34 ] proposed a constrained multi-objective optimization framework, learning a model that satisfies the fairness constraints of all clients with consistent performance by optimizing the agent’s maximal function involving all objectives. Li et al [ 35 ] proposed the Ditto method, employing inter-client fine-tuning to minimize individual losses significantly.…”

Section: Background and Related Workmentioning

confidence: 99%

Personalized Fair Split Learning for Resource-Constrained Internet of Things

Chen,

Peng

et al. 2023

Sensors

View full text Add to dashboard Cite

With the flourishing development of the Internet of Things (IoT), federated learning has garnered significant attention as a distributed learning method aimed at preserving the privacy of participant data. However, certain IoT devices, such as sensors, face challenges in effectively employing conventional federated learning approaches due to limited computational and storage resources, which hinder their ability to train complex local models. Additionally, in IoT environments, devices often face problems of data heterogeneity and uneven benefit distribution between them. To address these challenges, a personalized and fair split learning framework is proposed for resource-constrained clients. This framework first adopts a U-shaped structure, dividing the model to enable resource-constrained clients to offload subsets of the foundational model to a central server while retaining personalized model subsets locally to meet the specific personalized requirements of different clients. Furthermore, to ensure fair benefit distribution, a model-aggregation method with optimized aggregation weights is used. This method reasonably allocates model-aggregation weights based on the contributions of clients, thereby achieving collaborative fairness. Experimental results demonstrate that, in three distinct data heterogeneity scenarios, employing personalized training through this framework exhibits higher accuracy compared to existing baseline methods. Simultaneously, the framework ensures collaborative fairness, fostering a more balanced and sustainable cooperation among IoT devices.

show abstract

“…VIRTUAL (Corinzia et al, 2019) is a federated MTL framework for non-convex models based on a hierarchical Bayesian network formed by the central server and the clients, and inference is performed using variational methods. SPO (Cui et al, 2021) applies Specific Pareto Optimization to identify the optimal collaborator sets and learn a hypernetwork for all clients. While also aiming to identify necessary collaborators, SPO adopts a centralized FL setting with clients jointly training the hypernetwork.…”

Section: Model Regularization / Interpolationmentioning

confidence: 99%

Find Your Friends: Personalized Federated Learning with the Right Collaborators

Sui¹,

Wen²,

Lau³

et al. 2022

Preprint

View full text Add to dashboard Cite

In the traditional federated learning setting, a central server coordinates a network of clients to train one global model. However, the global model may serve many clients poorly due to data heterogeneity. Moreover, there may not exist a trusted central party that can coordinate the clients to ensure that each of them can benefit from others. To address these concerns, we present a novel decentralized framework, FedeRiCo, where each client can learn as much or as little from other clients as is optimal for its local data distribution. Based on expectationmaximization, FedeRiCo estimates the utilities of other participants' models on each client's data so that everyone can select the right collaborators for learning. As a result, our algorithm outperforms other federated, personalized, and/or decentralized approaches on several benchmark datasets, being the only approach that consistently performs better than training with local data only.

show abstract

Collaboration Equilibrium in Federated Learning

Cited by 11 publications

References 14 publications

Learning across diverse biomedical data modalities and cohorts: Challenges and opportunities for innovation

Learning across diverse biomedical data modalities and cohorts: Challenges and opportunities for innovation

Personalized Fair Split Learning for Resource-Constrained Internet of Things

Find Your Friends: Personalized Federated Learning with the Right Collaborators

Contact Info

Product

Resources

About