Tune As You Scale: Hyperparameter Optimization For Compute Efficient Training
Generally Intelligent（通用智能公司）；UC Berkeley（加州大学伯克利分校）
Abraham J. Fetterman, Ellie Kitanidis, Joshua Albrecht, Zachary Polizzi, Bryden Fogelman, Maksis Knutins, Bartosz Wróblewski, James B. Simon, Kanjun Qiu
深度学习模型的超参数调优可在相同计算量下带来数量级的性能提升。尽管如此，系统性调优并不常见，尤其是针对大模型——这类模型评估成本高昂，且往往包含大量超参数，需要在权衡取舍、预算和搜索范围之间做出艰难的判断决策。为解决这些问题并提出一种可稳健调优大模型的实用方法，我们提出了成本感知帕累托区域贝叶斯搜索（Cost-Aware Pareto Region Bayesian Search，CARBS），这是一种围绕性能-成本帕累托前沿进行局部搜索的贝叶斯优化算法。CARBS即使在超参数众多的无界搜索空间中也表现出色，它能学习缩放关系，从而在模型规模扩大时仍可完成调优，并将调优过程中大量的“黑魔法”式操作自动化。在我们的实验结果中，仅通过调优一个简单基线（原始ProcGen论文中提供的PPO），我们就有效攻克了整个ProcGen基准测试。我们还复现了Chinchilla项目（Hoffmann等人，2022）中的模型规模与训练token数的缩放结果，同时通过一套简便的自动化流程发现了所有其他超参数的缩放定律；该流程的计算消耗显著更低，且适用于任意深度学习问题（而非仅语言模型）。
英文摘要：
Hyperparameter tuning of deep learning models can lead to order-of-magnitude performance gains for the same amount of compute. Despite this, systematic tuning is uncommon, particularly for large models, which are expensive to evaluate and tend to have many hyperparameters, necessitating difficult judgment calls about tradeoffs, budgets, and search bounds. To address these issues and propose a practical method for robustly tuning large models, we present Cost-Aware Pareto Region Bayesian Search (CARBS), a Bayesian optimization algorithm that performs local search around the performance-cost Pareto frontier. CARBS does well even in unbounded search spaces with many hyperparameters, learns scaling relationships so that it can tune models even as they are scaled up, and automates much of the "black magic" of tuning. Among our results, we effectively solve the entire ProcGen benchmark just by tuning a simple baseline (PPO, as provided in the original ProcGen paper). We also reproduce the model size vs. training tokens scaling result from the Chinchilla project (Hoffmann et al. 2022), while simultaneously discovering scaling laws for every other hyperparameter, via an easy automated process that uses significantly less compute and is applicable to any deep learning problem (not just language models).
Computing Institute University of Campinas（坎皮纳斯大学计算研究所）；Facultad de Ciencias y Filosofía Universidad Peruana Cayetano Heredia（秘鲁卡耶塔诺埃雷迪亚大学科学与哲学学院）
SVM with an RBF kernel is usually one of the best classification algorithms for most data sets, but it is important to tune the two hyperparameters $C$ and $γ$ to the data itself. In general, the selection of the hyperparameters is a non-convex optimization problem and thus many algorithms have been proposed to solve it, among them: grid search, random search, Bayesian optimization, simulated annealing, particle swarm optimization, Nelder Mead, and others. There have also been proposals to decouple the selection of $γ$ and $C$. We empirically compare 18 of these proposed search algorithms (wit…
Deep learning has achieved impressive results on many problems. However, it requires high degree of expertise or a lot of experience to tune well the hyperparameters, and such manual tuning process is likely to be biased. Moreover, it is not practical to try out as many different hyperparameter configurations in deep learning as in other machine learning scenarios, because evaluating each single hyperparameter configuration in deep learning would mean training a deep neural network, which usually takes quite long time. Hyperband algorithm achieves state-of-the-art performance on various hyperp…
Bayesian Optimization is the state of the art technique for the optimization of black boxes, i.e., functions where we do not have access to their analytical expression nor its gradients, they are expensive to evaluate and its evaluation is noisy. The most popular application of bayesian optimization is the automatic hyperparameter tuning of machine learning algorithms, where we obtain the best configuration of machine learning algorithms by optimizing the estimation of the generalization error of these algorithms. Despite being applied with success, bayesian optimization methodologies also hav…
In this paper we develop a Bayesian optimization based hyperparameter tuning framework inspired by statistical learning theory for classifiers. We utilize two key facts from PAC learning theory; the generalization bound will be higher for a small subset of data compared to the whole, and the highest accuracy for a small subset of data can be achieved with a simple model. We initially tune the hyperparameters on a small subset of training data using Bayesian optimization. While tuning the hyperparameters on the whole training data, we leverage the insights from the learning theory to seek more …
Tuning hyperparameters of learning algorithms is hard because gradients are usually unavailable. We compute exact gradients of cross-validation performance with respect to all hyperparameters by chaining derivatives backwards through the entire training procedure. These gradients allow us to optimize thousands of hyperparameters, including step-size and momentum schedules, weight initialization distributions, richly parameterized regularization schemes, and neural network architectures. We compute hyperparameter gradients by exactly reversing the dynamics of stochastic gradient descent with mo…
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 nove…
Novel technologies in automated machine learning ease the complexity of algorithm selection and hyperparameter optimization. Hyperparameters are important for machine learning models as they significantly influence the performance of machine learning models. Many optimization techniques have achieved notable success in hyperparameter tuning and surpassed the performance of human experts. However, depending on such techniques as blackbox algorithms can leave machine learning practitioners without insight into the relative importance of different hyperparameters. In this paper, we consider build…
Amazon Web Services（亚马逊云科技）；Carnegie Mellon University（卡内基梅隆大学）
Given the increasing importance of machine learning (ML) in our lives, several algorithmic fairness techniques have been proposed to mitigate biases in the outcomes of the ML models. However, most of these techniques are specialized to cater to a single family of ML models and a specific definition of fairness, limiting their adaptibility in practice. We introduce a general constrained Bayesian optimization (BO) framework to optimize the performance of any ML model while enforcing one or multiple fairness constraints. BO is a model-agnostic optimization method that has been successfully applie…
Competence Center for High Performance Computing Fraunhofer ITWM Kaiserslautern（弗劳恩霍夫研究所ITWM高性能计算能力中心（凯泽斯劳滕））；Institute for Machine Learning and Analytics (IMLA) Offenburg University（奥芬堡大学机器学习与分析研究所（IMLA））
We introduce an open source python framework named PHS - Parallel Hyperparameter Search to enable hyperparameter optimization on numerous compute instances of any arbitrary python function. This is achieved with minimal modifications inside the target function. Possible applications appear in expensive to evaluate numerical computations which strongly depend on hyperparameters such as machine learning. Bayesian optimization is chosen as a sample efficient method to propose the next query set of parameters.
Multi-Objective Hyperparameter Optimization in Machine Learning -- An Overview
Fraunhofer Institut für integrierte Schaltungen（弗劳恩霍夫集成电路研究所）；Ludwig-Maximilians-Universität München（慕尼黑大学）；Technische Universität Dortmund（多特蒙德工业大学）；Universidad Pontificia Comillas（康普卢滕塞大学）；Warwick Business School（华威商学院）
For a large class of regularized models, leave-one-out cross-validation can be efficiently estimated with an approximate leave-one-out formula (ALO). We consider the problem of adjusting hyperparameters so as to optimize ALO. We derive efficient formulas to compute the gradient and hessian of ALO and show how to apply a second-order optimizer to find hyperparameters. We demonstrate the usefulness of the proposed approach by finding hyperparameters for regularized logistic regression and ridge regression on various real-world data sets.
Tuning the Tuner: Introducing Hyperparameter Optimization for Auto-Tuning
LIACS, Leiden University Netherlands eScience Center Leiden \& Amsterdam, the Netherlands ORCID: 0000-0003-2295-8263；LIACS, Leiden University Leiden, the Netherlands ORCID: 0000-0002-2947-9444；LIACS, Leiden University Netherlands eScience Center Leiden \& Amsterdam, the Netherlands ORCID: 0000-0002-7508-3272
Dana-Farber Cancer Institute（丹娜-法伯癌症研究所）；Harvard T.H. Chan School of Public Health（哈佛大学陈曾熙公共卫生学院）
Statistical learning models have been growing in popularity in recent years. Many of these models have hyperparameters that must be tuned for models to perform well. Tuning these parameters is not trivial. EZtune is an R package with a simple user interface that can tune support vector machines, adaboost, gradient boosting machines, and elastic net. We first provide a brief summary of the the models that EZtune can tune, including a discussion of each of their hyperparameters. We then compare the ease of using EZtune, caret, and tidymodels. This is followed with a comparison of the accuracy an…
Modern machine learning algorithms are increasingly computationally demanding, requiring specialized hardware and distributed computation to achieve high performance in a reasonable time frame. Many hyperparameter search algorithms have been proposed for improving the efficiency of model selection, however their adaptation to the distributed compute environment is often ad-hoc. We propose Tune, a unified framework for model selection and training that provides a narrow-waist interface between training scripts and search algorithms. We show that this interface meets the requirements for a broad…
Eindhoven University of Technology（埃因霍温理工大学）；Columbia University（哥伦比亚大学）
The performance of many machine learning algorithms depends on their hyperparameter settings. The goal of this study is to determine whether it is important to tune a hyperparameter or whether it can be safely set to a default value. We present a methodology to determine the importance of tuning a hyperparameter based on a non-inferiority test and tuning risk: the performance loss that is incurred when a hyperparameter is not tuned, but set to a default value. Because our methods require the notion of a default parameter, we present a simple procedure that can be used to determine reasonable d…
The random forest algorithm (RF) has several hyperparameters that have to be set by the user, e.g., the number of observations drawn randomly for each tree and whether they are drawn with or without replacement, the number of variables drawn randomly for each split, the splitting rule, the minimum number of samples that a node must contain and the number of trees. In this paper, we first provide a literature review on the parameters' influence on the prediction performance and on variable importance measures. It is well known that in most cases RF works reasonably well with the default values …
Department of Computer Science, University of Toronto（多伦多大学计算机科学系）
Machine learning models are often tuned by nesting optimization of model weights inside the optimization of hyperparameters. We give a method to collapse this nested optimization into joint stochastic optimization of weights and hyperparameters. Our process trains a neural network to output approximately optimal weights as a function of hyperparameters. We show that our technique converges to locally optimal weights and hyperparameters for sufficiently large hypernetworks. We compare this method to standard hyperparameter optimization strategies and demonstrate its effectiveness for tuning tho…
Network embedding (NE) can generate succinct node representations for massive-scale networks and enable direct applications of common machine learning methods to the network structure. Various NE algorithms have been proposed and used in a number of applications, such as node classification and link prediction. NE algorithms typically contain hyperparameters that are key to performance, but the hyperparameter tuning process can be time consuming. It is desirable to have the hyperparameters tuned within a specified length of time. Although AutoML methods have been applied to the hyperparameter …
Knowledge graph embeddings rank among the most successful methods for link prediction in knowledge graphs, i.e., the task of completing an incomplete collection of relational facts. A downside of these models is their strong sensitivity to model hyperparameters, in particular regularizers, which have to be extensively tuned to reach good performance [Kadlec et al., 2017]. We propose an efficient method for large scale hyperparameter tuning by interpreting these models in a probabilistic framework. After a model augmentation that introduces per-entity hyperparameters, we use a variational expec…
Model-agnostic meta-learning (MAML) is a meta-learning technique to train a model on a multitude of learning tasks in a way that primes the model for few-shot learning of new tasks. The MAML algorithm performs well on few-shot learning problems in classification, regression, and fine-tuning of policy gradients in reinforcement learning, but comes with the need for costly hyperparameter tuning for training stability. We address this shortcoming by introducing an extension to MAML, called Alpha MAML, to incorporate an online hyperparameter adaptation scheme that eliminates the need to tune meta-…
Conventional hyperparameter optimization methods are computationally intensive and hard to generalize to scenarios that require dynamically adapting hyperparameters, such as life-long learning. Here, we propose an online hyperparameter optimization algorithm that is asymptotically exact and computationally tractable, both theoretically and practically. Our framework takes advantage of the analogy between hyperparameter optimization and parameter learning in recurrent neural networks (RNNs). It adapts a well-studied family of online learning algorithms for RNNs to tune hyperparameters and netwo…
Federal Technology University（联邦技术大学）；University of São Paulo（圣保罗大学）；Universidade Estadual Paulista（保利斯塔州立大学）；Eindhoven University of Technology（埃因霍温理工大学）
For many machine learning algorithms, predictive performance is critically affected by the hyperparameter values used to train them. However, tuning these hyperparameters can come at a high computational cost, especially on larger datasets, while the tuned settings do not always significantly outperform the default values. This paper proposes a recommender system based on meta-learning to identify exactly when it is better to use default values and when to tune hyperparameters for each new dataset. Besides, an in-depth analysis is performed to understand what they take into account for their d…
Federal University of São Carlos（圣卡洛斯联邦大学）；São Paulo State University（圣保罗州立大学）
Deep learning techniques, such as Deep Boltzmann Machines (DBMs), have received considerable attention over the past years due to the outstanding results concerning a variable range of domains. One of the main shortcomings of these techniques involves the choice of their hyperparameters, since they have a significant impact on the final results. This work addresses the issue of fine-tuning hyperparameters of Deep Boltzmann Machines using metaheuristic optimization techniques with different backgrounds, such as swarm intelligence, memory- and evolutionary-based approaches. Experiments conducted…
In recent years, there have been many practical applications of anomaly detection such as in predictive maintenance, detection of credit fraud, network intrusion, and system failure. The goal of anomaly detection is to identify in the test data anomalous behaviors that are either rare or unseen in the training data. This is a common goal in predictive maintenance, which aims to forecast the imminent faults of an appliance given abundant samples of normal behaviors. Local outlier factor (LOF) is one of the state-of-the-art models used for anomaly detection, but the predictive performance of LOF…
IBM Germany（IBM德国）；Duale Hochschule Baden-Württemberg（巴登-符腾堡双元制大学）；IBM Research（IBM研究院）
In this short paper we investigate whether meta-learning techniques can be used to more effectively tune the hyperparameters of machine learning models using successive halving (SH). We propose a novel variant of the SH algorithm (MeSH), that uses meta-regressors to determine which candidate configurations should be eliminated at each round. We apply MeSH to the problem of tuning the hyperparameters of a gradient-boosted decision tree model. By training and tuning our meta-regressors using existing tuning jobs from 95 datasets, we demonstrate that MeSH can often find a superior solution to bot…
Alberta Machine Intelligence Institute (Amii)（阿尔伯塔机器智能研究所）；University of Alberta（阿尔伯塔大学）
The performance of reinforcement learning (RL) agents is sensitive to the choice of hyperparameters. In real-world settings like robotics or industrial control systems, however, testing different hyperparameter configurations directly on the environment can be financially prohibitive, dangerous, or time consuming. We propose a new approach to tune hyperparameters from offline logs of data, to fully specify the hyperparameters for an RL agent that learns online in the real world. The approach is conceptually simple: we first learn a model of the environment from the offline data, which we call …
While existing work on neural architecture search (NAS) tunes hyperparameters in a separate post-processing step, we demonstrate that architectural choices and other hyperparameter settings interact in a way that can render this separation suboptimal. Likewise, we demonstrate that the common practice of using very few epochs during the main NAS and much larger numbers of epochs during a post-processing step is inefficient due to little correlation in the relative rankings for these two training regimes. To combat both of these problems, we propose to use a recent combination of Bayesian optimi…
University of Toronto（多伦多大学）；Vector Institute（向量研究所）
We propose an algorithm for inexpensive gradient-based hyperparameter optimization that combines the implicit function theorem (IFT) with efficient inverse Hessian approximations. We present results about the relationship between the IFT and differentiating through optimization, motivating our algorithm. We use the proposed approach to train modern network architectures with millions of weights and millions of hyper-parameters. For example, we learn a data-augmentation network - where every weight is a hyperparameter tuned for validation performance - outputting augmented training examples. Jo…
The extent to which the arbitrarily selected L2 regularization hyperparameter value affects the outcome of semantic segmentation with deep learning is demonstrated. Demonstrations rely on training U-net on small LGE-MRI datasets using the arbitrarily selected L2 regularization values. The remaining hyperparameters are to be manually adjusted or tuned only when 10 % of all epochs are reached before the training validation accuracy reaches 90%. Semantic segmentation with deep learning outcomes are objectively and subjectively evaluated against the manual ground truth segmentation.
Bayesian optimization (BO) is a sample efficient approach to automatically tune the hyperparameters of machine learning models. In practice, one frequently has to solve similar hyperparameter tuning problems sequentially. For example, one might have to tune a type of neural network learned across a series of different classification problems. Recent work on multi-task BO exploits knowledge gained from previous tuning tasks to speed up a new tuning task. However, previous approaches do not account for the fact that BO is a sequential decision making procedure. Hence, there is in general a misma…
Deploying machine learning models requires high model quality and needs to comply with application constraints. That motivates hyperparameter optimization (HPO) to tune model configurations under deployment constraints. The constraints often require additional computation cost to evaluate, and training ineligible configurations can waste a large amount of tuning cost. In this work, we propose an Adaptive Constraint-aware Early stopping (ACE) method to incorporate constraint evaluation into trial pruning during HPO. To minimize the overall optimization cost, ACE estimates the cost-effective con…
Temperature is a widely used hyperparameter in various tasks involving neural networks, such as classification or metric learning, whose choice can have a direct impact on the model performance. Most of existing works select its value using hyperparameter optimization methods requiring several runs to find the optimal value. We propose to analyze the impact of temperature on classification tasks by describing a dataset as a set of statistics computed on representations on which we can build a heuristic giving us a default value of temperature. We study the correlation between these extracted s…
University of Freiburg（弗莱堡大学）；University of Technology Nuremberg（纽伦堡工业大学）
Accurate value estimates are important for off-policy reinforcement learning. Algorithms based on temporal difference learning typically are prone to an over- or underestimation bias building up over time. In this paper, we propose a general method called Adaptively Calibrated Critics (ACC) that uses the most recent high variance but unbiased on-policy rollouts to alleviate the bias of the low variance temporal difference targets. We apply ACC to Truncated Quantile Critics, which is an algorithm for continuous control that allows regulation of the bias with a hyperparameter tuned per environme…
