Automatic model training under restrictive time constraints
From MaRDI portal
Publication:2108929
Abstract: We develop a hyperparameter optimisation algorithm, Automated Budget Constrained Training (AutoBCT), which balances the quality of a model with the computational cost required to tune it. The relationship between hyperparameters, model quality and computational cost must be learnt and this learning is incorporated directly into the optimisation problem. At each training epoch, the algorithm decides whether to terminate or continue training, and, in the latter case, what values of hyperparameters to use. This decision weighs optimally potential improvements in the quality with the additional training time and the uncertainty about the learnt quantities. The performance of our algorithm is verified on a number of machine learning problems encompassing random forests and neural networks. Our approach is rooted in the theory of Markov decision processes with partial information and we develop a numerical method to compute the value function and an optimal strategy.
Recommendations
- Hyperband: a novel bandit-based approach to hyperparameter optimization
- Fast Bayesian hyperparameter optimization on large datasets
- A general framework for constrained Bayesian optimization using information-based search
- scientific article; zbMATH DE number 6276119
- Hyperparameter optimization in learning systems
Cites work
- scientific article; zbMATH DE number 3144846 (Why is no real title available?)
- scientific article; zbMATH DE number 700016 (Why is no real title available?)
- A numerical algorithm for fully nonlinear HJB equations: an approach by control randomization
- Algorithm runtime prediction: methods \& evaluation
- Comparison of least squares Monte Carlo methods with applications to energy real options
- Deep neural networks algorithms for stochastic control problems on finite horizon: convergence analysis
- Deep neural networks algorithms for stochastic control problems on finite horizon: numerical applications
- Estimation and control of dynamical systems
- Generating random correlation matrices based on vines and extended onion method
- Hyperband: a novel bandit-based approach to hyperparameter optimization
- Regularization and Variable Selection Via the Elastic Net
- Stochastic optimal control. The discrete time case
- Valuing American options by simulation: a simple least-squares approach
This page was built for publication: Automatic model training under restrictive time constraints
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2108929)
