skip to main content

Title: Federated Hyperparameter Tuning: Challenges, Baselines, and Connections to Weight-Sharing
Tuning hyperparameters is a crucial but arduous part of the machine learning pipeline. Hyperparameter optimization is even more challenging in federated learning, where models are learned over a distributed network of heterogeneous devices; here, the need to keep data on device and perform local training makes it difficult to efficiently train and evaluate configurations. In this work, we investigate the problem of federated hyperparameter tuning. We first identify key challenges and show how standard approaches may be adapted to form baselines for the federated setting. Then, by making a novel connection to the neural architecture search technique of weight-sharing, we introduce a new method, FedEx, to accelerate federated hyperparameter tuning that is applicable to widely-used federated optimization methods such as FedAvg and recent variants. Theoretically, we show that a FedEx variant correctly tunes the on-device learning rate in the setting of online convex optimization across devices. Empirically, we show that FedEx can outperform natural baselines for federated hyperparameter tuning by several percentage points on the Shakespeare, FEMNIST, and CIFAR-10 benchmarks, obtaining higher accuracy using the same training budget.
; ; ; ; ; ;
Award ID(s):
Publication Date:
Journal Name:
Advances in neural information processing systems
Sponsoring Org:
National Science Foundation
More Like this
  1. Solving a bilevel optimization problem is at the core of several machine learning problems such as hyperparameter tuning, data denoising, meta- and few-shot learning, and training-data poisoning. Different from simultaneous or multi-objective optimization, the steepest descent direction for minimizing the upper-level cost in a bilevel problem requires the inverse of the Hessian of the lower-level cost. In this work, we propose a novel algorithm for solving bilevel optimization problems based on the classical penalty function approach. Our method avoids computing the Hessian inverse and can handle constrained bilevel problems easily. We prove the convergence of the method under mild conditionsmore »and show that the exact hypergradient is obtained asymptotically. Our method's simplicity and small space and time complexities enable us to effectively solve large-scale bilevel problems involving deep neural networks. We present results on data denoising, few-shot learning, and training-data poisoning problems in a large-scale setting. Our results show that our approach outperforms or is comparable to previously proposed methods based on automatic differentiation and approximate inversion in terms of accuracy, run-time, and convergence speed.« less
  2. Edge Computing (EC) has seen a continuous rise in its popularity as it provides a solution to the latency and communication issues associated with edge devices transferring data to remote servers. EC achieves this by bringing the cloud closer to edge devices. Even though EC does an excellent job of solving the latency and communication issues, it does not solve the privacy issues associated with users transferring personal data to the nearby edge server. Federated Learning (FL) is an approach that was introduced to solve the privacy issues associated with data transfers to distant servers. FL attempts to resolve thismore »issue by bringing the code to the data, which goes against the traditional way of sending the data to remote servers. In FL, the data stays on the source device, and a Machine Learning (ML) model used to train the local data is brought to the end device instead. End devices train the ML model using local data and then send the model updates back to the server for aggregation. However, this process of asking random devices to train a model using its local data has potential risks such as a participant poisoning the model using malicious data for training to produce bogus parameters. In this paper, an approach to mitigate data poisoning attacks in a federated learning setting is investigated. The application of the approach is highlighted, and the practical and secure nature of this approach is illustrated as well using numerical results.« less
  3. Hyperparameter tuning is a necessary step in training and deploying machine learning models. Most prior work on hyperparameter tuning has studied methods for maximizing model accuracy under a time constraint, assuming a fixed cluster size. While this is appropriate in data center environments, the increased deployment of machine learning workloads in cloud settings necessitates studying hyperparameter tuning with an elastic cluster size and time and monetary budgets. While recent work has leveraged the elasticity of the cloud to minimize the execution cost of a pre-determined hyperparameter tuning job originally designed for fixed-cluster sizes, they do not aim to maximize accuracy.more »In this work, we aim to maximize accuracy given time and cost constraints. We introduce SEER---Sequential Elimination with Elastic Resources, an algorithm that tests different hyperparameter values in the beginning and maintains varying degrees of parallelism among the promising configurations to ensure that they are trained sufficiently before the deadline. Unlike fixed cluster size methods, it is able to exploit the flexibility in resource allocation the elastic setting has to offer in order to avoid undesirable effects of sublinear scaling. Furthermore, SEER can be easily integrated into existing systems and makes minimal assumptions about the workload. On a suite of benchmarks, we demonstrate that SEER outperforms both existing methods for hyperparameter tuning on a fixed cluster as well as naive extensions of these algorithms to the cloud setting.« less
  4. In this work, we show that for a nontrivial hypothesis class C, we can estimate the distance of a target function f to C (estimate the error rate of the best h∈C) using substantially fewer labeled examples than would be needed to actually {\em learn} a good h∈C. Specifically, we show that for the class C of unions of d intervals on the line, in the active learning setting in which we have access to a pool of unlabeled examples drawn from an arbitrary underlying distribution D, we can estimate the error rate of the best h∈C to an additivemore »error ϵ with a number of label requests that is {\em independent of d} and depends only on ϵ. In particular, we make O((1/ϵ^6)log(1/ϵ)) label queries to an unlabeled pool of size O((d/ϵ^2)log(1/ϵ)). This task of estimating the distance of an unknown f to a given class C is called {\em tolerant testing} or {\em distance estimation} in the testing literature, usually studied in a membership query model and with respect to the uniform distribution. Our work extends that of Balcan et al. (2012) who solved the {\em non}-tolerant testing problem for this class (distinguishing the zero-error case from the case that the best hypothesis in the class has error greater than ϵ). We also consider the related problem of estimating the performance of a given learning algorithm A in this setting. That is, given a large pool of unlabeled examples drawn from distribution D, can we, from only a few label queries, estimate how well A would perform if the entire dataset were labeled and given as training data to A? We focus on k-Nearest Neighbor style algorithms, and also show how our results can be applied to the problem of hyperparameter tuning (selecting the best value of k for the given learning problem).« less
  5. We give a simple, fast algorithm for hyperparameter optimization inspired by techniques from the analysis of Boolean functions. We focus on the high-dimensional regime where the canonical example is training a neural network with a large number of hyperparameters. The algorithm --- an iterative application of compressed sensing techniques for orthogonal polynomials --- requires only uniform sampling of the hyperparameters and is thus easily parallelizable. Experiments for training deep neural networks on Cifar-10 show that compared to state-of-the-art tools (e.g., Hyperband and Spearmint), our algorithm finds significantly improved solutions, in some cases better than what is attainable by hand-tuning. Inmore »terms of overall running time (i.e., time required to sample various settings of hyperparameters plus additional computation time), we are at least an order of magnitude faster than Hyperband and Bayesian Optimization. We also outperform Random Search 8x. Additionally, our method comes with provable guarantees and yields the first improvements on the sample complexity of learning decision trees in over two decades. In particular, we obtain the first quasi-polynomial time algorithm for learning noisy decision trees with polynomial sample complexity.« less