An Asymptotically Optimal Multi-Armed Bandit Algorithm and Hyperparameter Optimization

By Yimin Huang, Yujun Li, Zhenguo Li, Zhihua Zhang. 2020

8 citations recorded by the source corpus.

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).

Subfields: Artificial Intelligence, Management Science and Operations Research

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Zhenguo Li
2. Hang Xu
3. Wei Zhang
4. Lanqing Hong
5. Ning Kang
6. Xinyue Cai
7. Yong Li
8. Jiajin Zhang

In Management Science and Operations Research

1. Ning Kang
2. Wei Zhang
3. Hang Xu
4. Tong Zhang
5. Zhenguo Li
6. Julien Mairal
7. Zeyuan Allen-Zhu
8. Lin Xiao

Citations in the corpus

• VEGA: Towards an End-to-End Configurable AutoML Pipeline (2020, methodology): “…optimization and bandit-based methods, in order to achieve the best of both worlds, and proposes a new practical state-of-the-art hyperparameter optimization method, which consistently outperforms both Bayesian optimization and Hyperband on a wide range of problem types BOSS (Huang et al., 2020).”