Population Based Training

GeneralIntroduced 200018 papers
Source Paper

Description

Population Based Training, or PBT, is an optimization method for finding parameters and hyperparameters, and extends upon parallel search methods and sequential optimisation methods. It leverages information sharing across a population of concurrently running optimisation processes, and allows for online propagation/transfer of parameters and hyperparameters between members of the population based on their performance. Furthermore, unlike most other adaptation schemes, the method is capable of performing online adaptation of hyperparameters -- which can be particularly important in problems with highly non-stationary learning dynamics, such as reinforcement learning settings. PBT is decentralised and asynchronous, although it could also be executed semi-serially or with partial synchrony if there is a binding budget constraint.

Papers Using This Method

Shrink-Perturb Improves Architecture Mixing during Population Based Training for Neural Architecture Search2023-07-28Multi-Objective Population Based Training2023-06-02Reducing Exploitability with Population Based Training2022-08-10Faster Improvement Rate Population Based Training2021-09-28Automated Graph Learning via Population Based Self-Tuning GCN2021-07-09Tuning Mixed Input Hyperparameters on the Fly for Efficient Population Based AutoRL2021-06-30Training Learned Optimizers with Randomly Initialized Learned Optimizers2021-01-14Population Based Training for Data Augmentation and Regularization in Speech Recognition2020-10-08Improving 3D Object Detection through Progressive Population Based Augmentation2020-04-02Accelerating and Improving AlphaZero Using Population Based Training2020-03-13DADA: Differentiable Automatic Data Augmentation2020-03-08Data Efficient Training for Reinforcement Learning with Adaptive Behavior Policy Sharing2020-02-12Provably Efficient Online Hyperparameter Optimization with Population-Based Bandits2020-02-06Fast Efficient Hyperparameter Tuning for Policy Gradient Methods2019-12-01Population Based Augmentation: Efficient Learning of Augmentation Policy Schedules2019-05-14Semi-supervised and Population Based Training for Voice Commands Recognition2019-05-10Fast Efficient Hyperparameter Tuning for Policy Gradients2019-02-18Population Based Training of Neural Networks2017-11-27