An Asymptotically Optimal Multi-Armed Bandit Algorithm and Hyperparameter Optimization
arXiv:2007.05670
Abstract
The evaluation of hyperparameters, neural architectures, or data augmentation policies becomes a critical model selection problem in advanced deep learning with a large hyperparameter search space. In this paper, we propose an efficient and robust bandit-based algorithm called Sub-Sampling (SS) in the scenario of hyperparameter search evaluation. It evaluates the potential of hyperparameters by the sub-samples of observations and is theoretically proved to be optimal under the criterion of cumulative regret. We further combine SS with Bayesian Optimization and develop a novel hyperparameter optimization algorithm called BOSS. Empirical studies validate our theoretical arguments of SS and demonstrate the superior performance of BOSS on a number of applications, including Neural Architecture Search (NAS), Data Augmentation (DA), Object Detection (OD), and Reinforcement Learning (RL).
References in corpus (24)
- Practical Bayesian Optimization of Machine Learning Algorithms
- ADADELTA: An Adaptive Learning Rate Method
- R-FCN: Object Detection via Region-based Fully Convolutional Networks
- Trust Region Policy Optimization
- Improved Regularization of Convolutional Neural Networks with Cutout
- Wide Residual Networks
- Neural Architecture Search: A Survey
- DARTS: Differentiable Architecture Search
- AutoAugment: Learning Augmentation Policies from Data
- Scalable Bayesian Optimization Using Deep Neural Networks
- Predictive Entropy Search for Efficient Global Optimization of Black-box Functions
- Understanding and Robustifying Differentiable Architecture Search
- DARTS+: Improved Differentiable Architecture Search with Early Stopping
- Batch Bayesian Optimization via Local Penalization
- Max-value Entropy Search for Efficient Bayesian Optimization
- MetaAnchor: Learning to Detect Objects with Customized Anchors
- Rethinking the Hyperparameters for Fine-tuning
- Combination of Hyperband and Bayesian Optimization for Hyperparameter Optimization in Deep Learning
- A Survey of Online Experiment Design with the Stochastic Multi-Armed Bandit
- Learning search spaces for Bayesian optimization: Another view of hyperparameter transfer learning
- Enabling hyperparameter optimization in sequential autoencoders for spiking neural data
- TuPAQ: An Efficient Planner for Large-scale Predictive Analytic Queries
- Hyperparameter Learning via Distributional Transfer
- Oracle inequalities for computationally adaptive model selection