An official website of the United States government
We consider a distributed empirical risk minimization (ERM) optimization problem with communication efficiency and privacy requirements, motivated by the federated learn- ing (FL) framework. We propose a distributed communication-efficient and local differentially private stochastic gradient descent (CLDP-SGD) algorithm and analyze its communication, privacy, and convergence trade-offs. Since each iteration of the CLDP- SGD aggregates the client-side local gradients, we develop (optimal) communication-efficient schemes for mean estimation for several lp spaces under local differential privacy (LDP). To overcome performance limitation of LDP, CLDP-SGD takes advantage of the inherent privacy amplification provided by client sub- sampling and data subsampling at each se- lected client (through SGD) as well as the recently developed shuffled model of privacy. For convex loss functions, we prove that the proposed CLDP-SGD algorithm matches the known lower bounds on the centralized private ERM while using a finite number of bits per iteration for each client, i.e., effectively get- ting communication efficiency for “free”. We also provide preliminary experimental results supporting the theory.
more »
« less
Differentially Private Federated Learning with Shuffling and Client Self-Sampling
This paper studies a distributed optimization problem in the federated learning (FL) framework under differential privacy constraints, whereby a set of clients having local samples are connected to an untrusted server, who wants to learn a global model while preserving the privacy of clients’ local datasets. We propose a new client sampling called self-sampling that reflects the random availability of clients in the learning process in FL. We analyze the differential privacy of the SGD with client self-sampling by composing amplification by sub-sampling along with amplification by shuffling. Furthermore, we analyze the convergence of the proposed SGD algorithm showing that we can get a reasonable learning performance while preserving the privacy of clients’ data even with client self-sampling.
more »
« less
- PAR ID:
- 10280813
- Date Published:
- Journal Name:
- IEEE International Symposium on Information Theory (ISIT)
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Federated Averaging (FedAvg) and its variants are the most popular optimization algorithms in federated learning (FL). Previous convergence analyses of FedAvg either assume full client participation or partial client participation where the clients can be uniformly sampled. However, in practical cross-device FL systems, only a subset of clients that satisfy local criteria such as battery status, network connectivity, and maximum participation frequency requirements (to ensure privacy) are available for training at a given time. As a result, client availability follows a natural cyclic pattern. We provide (to our knowledge) the first theoretical framework to analyze the convergence of FedAvg with cyclic client participation with several different client optimizers such as GD, SGD, and shuffled SGD. Our analysis discovers that cyclic client participation can achieve a faster asymptotic convergence rate than vanilla FedAvg with uniform client participation under suitable conditions, providing valuable insights into the design of client sampling protocols.more » « less
-
Providing privacy protection has been one of the primary motivations of Federated Learning (FL). Recently, there has been a line of work on incorporating the formal privacy notion of differential privacy with FL. To guarantee the client-level differential privacy in FL algorithms, the clients’ transmitted model updates have to be clipped before adding privacy noise. Such clipping operation is substantially different from its counterpart of gradient clipping in the centralized differentially private SGD and has not been well-understood. In this paper, we first empirically demonstrate that the clipped FedAvg can perform surprisingly well even with substantial data heterogeneity when training neural networks, which is partly because the clients’ updates become similar for several popular deep architectures. Based on this key observation, we provide the convergence analysis of a differential private (DP) FedAvg algorithm and highlight the relationship between clipping bias and the distribution of the clients’ updates. To the best of our knowledge, this is the first work that rigorously investigates theoretical and empirical issues regarding the clipping operation in FL algorithms.more » « less
-
Federated Learning (FL) is a privacy-preserving distributed machine learning technique that enables individual clients (e.g., user participants, edge devices, or organizations) to train a model on their local data in a secure environment and then share the trained model with an aggregator to build a global model collaboratively. In this work, we propose FedDefender, a defense mechanism against targeted poisoning attacks in FL by leveraging differential testing. FedDefender first applies differential testing on clients’ models using a synthetic input. Instead of comparing the output (predicted label), which is unavailable for synthetic input, FedDefender fingerprints the neuron activations of clients’ models to identify a potentially malicious client containing a backdoor. We evaluate FedDefender using MNIST and FashionMNIST datasets with 20 and 30 clients, and our results demonstrate that FedDefender effectively mitigates such attacks, reducing the attack success rate (ASR) to 10% without deteriorating the global model performance.more » « less
-
Federated Learning (FL) has emerged as an effective paradigm for distributed learning systems owing to its strong potential in exploiting underlying data characteristics while preserving data privacy. In cases of practical data heterogeneity among FL clients in many Internet-of-Things (IoT) applications over wireless networks, however, existing FL frameworks still face challenges in capturing the overall feature properties of local client data that often exhibit disparate distributions. One approach is to apply generative adversarial networks (GANs) in FL to address data heterogeneity by integrating GANs to regenerate anonymous training data without exposing original client data to possible eavesdropping. Despite some successes, existing GAN-based FL frameworks still incur high communication costs and elicit other privacy concerns, limiting their practical applications. To this end, this work proposes a novel FL framework that only applies partial GAN model sharing. This new PS-FedGAN framework effectively addresses heterogeneous data distributions across clients and strengthens privacy preservation at reduced communication costs, especially over wireless networks. Our analysis demonstrates the convergence and privacy benefits of the proposed PS-FEdGAN framework. Through experimental results based on several well-known benchmark datasets, our proposed PS-FedGAN demonstrates strong potential to tackle FL under heterogeneous (non-IID) client data distributions, while improving data privacy and lowering communication overhead.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
