Federated learning (FL) enables collaborative learning across multiple clients. In most FL work, all clients train a single learning task. However, the recent proliferation of FL applications may increasingly require multiple FL tasks to be trained simultaneously, sharing clients’ computing resources, which we call Multiple-Model Federated Learning (MMFL). Current MMFL algorithms use naïve average-based client-task allocation schemes that often lead to unfair performance when FL tasks have heterogeneous difficulty levels, as the more difficult tasks may need more client participation to train effectively. Furthermore, in the MMFL setting, we face a further challenge that some clients may prefer training specific tasks to others, and may not even be willing to train other tasks, e.g., due to high computational costs, which may exacerbate unfairness in training outcomes across tasks. We address both challenges by firstly designing FedFairMMFL, a difficulty-aware algorithm that dynamically allocates clients to tasks in each training round, based on the tasks’ current performance levels. We provide guarantees on the resulting task fairness and FedFairMMFL’s convergence rate. We then propose novel auction designs that incentivizes clients to train multiple tasks, so as to fairly distribute clients’ training efforts across the tasks, and extend our convergence guarantees to this setting. We finally evaluate our algorithm with multiple sets of learning tasks on real world datasets, showing that our algorithm improves fairness by improving the final model accuracy and convergence speed of the worst performing tasks, while maintaining the average accuracy across tasks.

Federated learning allows distributed clients to train a shared machine learning model while preserving user privacy. In this framework, user devices (i.e., clients) perform local iterations of the learning algorithm on their data. These updates are periodically aggregated to form a shared model. Thus, a client represents the bundle of the user data, the device, and the user’s willingness to participate: since participating in federated learning requires clients to expend resources and reveal some information about their data, users may require some form of compensation to contribute to the training process. Recruiting more users generally results in higher accuracy, but slower completion time and higher cost. We propose the first work to theoretically analyze the resulting performance tradeoffs in deciding which clients to recruit for the federated learning algorithm. Our framework accounts for both accuracy (training and testing) and efficiency (completion time and cost) metrics. We provide solutions to this NP-Hard optimization problem and verify the value of client recruitment in experiments on synthetic and real-world data. The results of this work can serve as a guideline for the real-world deployment of federated learning and an initial investigation of the client recruitment problem.

Federated learning (FL) is an increasingly popular form of distributed learning across devices such as sensors and smartphones. To amortize the effort and cost of setting up FL training in real world systems, in practice multiple machine learning tasks may be trained during one FL execution. However, given that the tasks have varying complexities, naïve methods of allocating resource-constrained devices to work on each task may lead to highly variable performance across the tasks. We instead propose an α -fair based allocation algorithm that dynamically allocates tasks to users during multi-model FL training, based on the prevailing loss levels.

Traditionally, clustered federated learning groups clients with the same data distribution into a cluster, so that every client is uniquely associated with one data distribution and helps train a model for this distribution. We relax this hard association assumption to soft clustered federated learning, which allows every local dataset to follow a mixture of multiple source distributions. We propose FedSoft, which trains both locally personalized models and high-quality cluster models in this setting. FedSoft limits client workload by using proximal updates to require the completion of only one optimization task from a subset of clients in every communication round. We show, analytically and empirically, that FedSoft effectively exploits similarities between the source distributions to learn personalized and cluster models that perform well.

Traditional federated learning algorithms impose strict requirements on the participation rates of devices, which limit the potential reach of federated learning. This paper extends the current learning paradigm to include devices that may become inactive, compute incomplete updates, and depart or arrive in the middle of training. We derive analytical results to illustrate how allowing more flexible device participation can affect the learning convergence when data is not independently and identically distributed (non-IID). We then propose a new federated aggregation scheme that converges even when devices may be inactive or return incomplete updates. We also study how the learning process can adapt to early departures or late arrivals, and analyze their impacts on the convergence.

Federated learning allows distributed clients to train a shared machine learning model while preserving user privacy. In this framework, an operator recruits user devices (i.e., clients) to occasionally perform local iterations of the learning algorithm on their data. We propose the first work to theoretically analyze the resulting performance tradeoffs in deciding which clients to recruit for federated learning, complementing other works on the selection of recruited clients in each iteration. Specifically, we define and optimize the tradeoffs between both accuracy (training and testing) and efficiency (completion time and cost) metrics. We provide efficient solutions to this NP-Hard optimization problem, and verify the value of client recruitment in experiments on synthetic and real-world data. The results of this work can serve as guidelines for the real-world deployment of federated learning and an initial investigation of the client recruitment problem.

Fog computing promises to enable machine learning tasks to scale to large amounts of data by distributing processing across connected devices. Two key challenges to achieving this goal are (i) heterogeneity in devices’ compute resources and (ii) topology constraints on which devices communicate with each other. We address these challenges by developing a novel network-aware distributed learning methodology where devices optimally share local data processing and send their learnt parameters to a server for periodic aggregation. Unlike traditional federated learning, our method enables devices to offload their data processing tasks to each other, with these decisions optimized to trade off costs associated with data processing, offloading, and discarding. We analytically characterize the optimal data transfer solution under different assumptions on the fog network scenario, showing for example that the value of offloading is approximately linear in the range of computing costs in the network when the cost of discarding is modeled as decreasing linearly in the amount of data processed at each node. Our experiments on real-world data traces from our testbed confirm that our algorithms improve network resource utilization substantially without sacrificing the accuracy of the learned model, for varying distributions of data across devices. We also investigate the effect of network dynamics on model learning and resource costs.

Recent growth in user demand for mobile data has strained mobile network infrastructure. One possible solution is to use mobile (i.e., moving) devices to supplement existing infrastructure according to users' needs at different times and locations. For instance, vehicles can be used as communication relays or computation points. However, it is unclear how much value these devices add relative to their deployment costs: they may, for instance, interfere with existing network infrastructure, limiting the potential benefits. We take the first step towards quantifying the value of this supplemental infrastructure by examining the use case of mobile caches. We consider a network operator using both mobile (e.g., vehicular) and stationary (small cell) caches, and find the optimal amount of both types of caches under time- and location-varying user demands, as a function of the cache prices. In doing so, we account for interference between users' connections to the different caches, which requires solving a non-convex optimization problem. We show that there exists a threshold price above which no vehicular caches are purchased. Moreover, as the network operator's budget increases, vehicular caching yields little additional value beyond that provided by small cell caches. These results may help network operators and cache providers find conditions under which vehicles add value to existing networks.

Fog computing promises to enable machine learning tasks to scale to large amounts of data by distributing processing across connected devices. Two key challenges to achieving this are (i) heterogeneity in devices' compute resources and (ii) topology constraints on which devices can communicate. We are the first to address these challenges by developing a network-aware distributed learning optimization methodology where devices process data for a task locally and send their learnt parameters to a server for aggregation at certain time intervals. Unlike traditional federated learning frameworks, our method enables devices to offload their data processing tasks, with these decisions determined through a convex data transfer optimization problem that trades off costs associated with devices processing, offloading, and discarding data points. We analytically characterize the optimal data transfer solution for different fog network topologies, showing for example that the value of a device offloading is approximately linear in the range of computing costs in the network. Our subsequent experiments on both synthetic and real-world datasets we collect confirm that our algorithms are able to improve network resource utilization substantially without sacrificing the accuracy of the learned model.