Shahryar Zehtabi

Shahryar Zehtabi, Seyyedali Hosseinalipour, Christopher G. Brinton

A recent emphasis of distributed learning research has been on federated learning (FL), in which model training is conducted by the data-collecting devices. Existing research on FL has mostly focused on a star topology learning architecture with synchronized (time-triggered) model training rounds, where the local models of the devices are periodically aggregated by a centralized coordinating node. However, in many settings, such a coordinating node may not exist, motivating efforts to fully decentralize FL. In this work, we propose a novel methodology for distributed model aggregations via asynchronous, event-triggered consensus iterations over the network graph topology. We consider heterogeneous communication event thresholds at each device that weigh the change in local model parameters against the available local resources in deciding the benefit of aggregations at each iteration. Through theoretical analysis, we demonstrate that our methodology achieves asymptotic convergence to the globally optimal learning model under standard assumptions in distributed learning and graph consensus literature, and without restrictive connectivity requirements on the underlying topology. Subsequent numerical results demonstrate that our methodology obtains substantial improvements in communication requirements compared with FL baselines.

Shahryar Zehtabi, Seyyedali Hosseinalipour, Christopher G. Brinton

Federated learning (FL) is a popular technique for distributing machine learning (ML) across a set of edge devices. In this paper, we study fully decentralized FL, where in addition to devices conducting training locally, they carry out model aggregations via cooperative consensus formation over device-to-device (D2D) networks. We introduce asynchronous, event-triggered communications among the devices to handle settings where access to a central server is not feasible. To account for the inherent resource heterogeneity and statistical diversity challenges in FL, we define personalized communication triggering conditions at each device that weigh the change in local model parameters against the available local network resources. We theoretically recover the $O(\ln{k} / \sqrt{k})$ convergence rate to the globally optimal model of decentralized gradient descent (DGD) methods in the setup of our methodology. We provide our convergence guarantees for the last iterates of models, under relaxed graph connectivity and data heterogeneity assumptions compared with the existing literature. To do so, we demonstrate a $B$-connected information flow guarantee in the presence of sporadic communications over the time-varying D2D graph. Our subsequent numerical evaluations demonstrate that our methodology obtains substantial improvements in convergence speed and/or communication savings compared to existing decentralized FL baselines.

Shahryar Zehtabi, Dong-Jun Han, Rohit Parasnis et al.

Decentralized federated learning (DFL) captures FL settings where both (i) model updates and (ii) model aggregations are exclusively carried out by the clients without a central server. Existing DFL works have mostly focused on settings where clients conduct a fixed number of local updates between local model exchanges, overlooking heterogeneity and dynamics in communication and computation capabilities. In this work, we propose Decentralized Sporadic Federated Learning ($\texttt{DSpodFL}$), a DFL methodology built on a generalized notion of $\textit{sporadicity}$ in both local gradient and aggregation processes. $\texttt{DSpodFL}$ subsumes many existing decentralized optimization methods under a unified algorithmic framework by modeling the per-iteration (i) occurrence of gradient descent at each client and (ii) exchange of models between client pairs as arbitrary indicator random variables, thus capturing $\textit{heterogeneous and time-varying}$ computation/communication scenarios. We analytically characterize the convergence behavior of $\texttt{DSpodFL}$ for both convex and non-convex models and for both constant and diminishing learning rates, under mild assumptions on the communication graph connectivity, data heterogeneity across clients, and gradient noises. We show how our bounds recover existing results from decentralized gradient descent as special cases. Experiments demonstrate that $\texttt{DSpodFL}$ consistently achieves improved training speeds compared with baselines under various system settings.

Marijan Fofonjka, Shahryar Zehtabi, Alireza Behtash et al.

When existing retrieval-augmented generation (RAG) solutions are intended to be used for new knowledge domains, it is necessary to update their encoders, which are taken to be pretrained large language models (LLMs). However, fully finetuning these large models is compute- and memory-intensive, and even infeasible when deployed on resource-constrained edge devices. We propose a novel encoder architecture in this work that addresses this limitation by using a frozen small language model (SLM), which satisfies the memory constraints of edge devices, and inserting a small adapter network before the transformer blocks of the SLM. The trainable adapter takes the token embeddings of the new corpus and learns to produce enhanced soft embeddings for it, while requiring significantly less compute power to update than full fine-tuning. We further propose a novel retrieval mechanism by attaching a classifier head to the SLM encoder, which is trained to learn a similarity mapping of the input embeddings to their corresponding documents. Finally, to enable the online fine-tuning of both (i) the encoder soft embeddings and (ii) the classifier-as-retriever on edge devices, we adopt federated learning (FL) and differential privacy (DP) to achieve an efficient, privacy-preserving, and product-grade training solution. We conduct a theoretical analysis of our methodology, establishing convergence guarantees under mild assumptions on gradient variance when deployed for general smooth nonconvex loss functions. Through extensive numerical experiments, we demonstrate (i) the efficacy of obtaining soft embeddings to enhance the encoder, (ii) training a classifier to improve the retriever, and (iii) the role of FL in achieving speedup.

Shahryar Zehtabi, Dong-Jun Han, Seyyedali Hosseinalipour et al.

Much of federated learning (FL) focuses on settings where local dataset statistics remain the same between training and testing. However, this assumption often does not hold in practice due to distribution shifts, motivating the development of domain generalization (DG) approaches that leverage source domain data to train models capable of generalizing to unseen target domains. In this paper, we are motivated by two major gaps in existing work on FL and DG: (1) the lack of formal mathematical analysis of DG objectives; and (2) DG research in FL being limited to the star-topology architecture. We develop Decentralized Federated Domain Generalization with Style Sharing ($\textit{StyleDDG}$), a decentralized DG algorithm which allows devices in a peer-to-peer network to achieve DG based on sharing style information inferred from their datasets. Additionally, we provide the first systematic approach to analyzing style-based DG training in decentralized networks. We cast existing centralized DG algorithms within our framework, and employ their formalisms to model $\textit{StyleDDG}$. We then obtain analytical conditions under which convergence of $\textit{StyleDDG}$ can be guaranteed. Through experiments on popular DG datasets, we demonstrate that $\textit{StyleDDG}$ can obtain significant improvements in accuracy across target domains with minimal communication overhead compared to baseline decentralized gradient methods.