Open AccessProceedings Article
Using machine learning to improve stochastic optimization
David H. Wolpert,Dev G. Rajnarayan +1 more
- 01 Jan 2013
- pp 146-148
TL;DR: It is shown how machine learning provides more principled alternatives to (adaptively) set that hyperparameter, and it is demonstrated that these alternatives can substantially improve optimization performance.
read more
Abstract: In many stochastic optimization algorithms there is a hyperparameter that controls how the next sampling distribution is determined from the current data set of samples of the objective function. This hyperparameter controls the exploration /exploitation trade-off of the next sample. Typically heuristic "rules of thumb" are used to set that hyperparameter, e.g., a pre-fixed annealing schedule. We show how machine learning provides more principled alternatives to (adaptively) set that hyperparameter, and demonstrate that these alternatives can substantially improve optimization performance.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
•Posted Content
The Implications of the No-Free-Lunch Theorems for Meta-induction.
TL;DR: The No-Free-Lunch Theorem (NFL) as discussed by the authors states that there are as many priors for which any induction algorithm out-generalizes some induction algorithm as vice-versa.
11
Comprehensive characterization of the behaviors of estimation of distribution algorithms
TL;DR: This paper substantially extends previous work in the proposal of a taxonomy of problems for univariate EDAs, mainly by generalizing those results to EDAs that are able to deal with multivariate dependences among the variables of the problem.
7
Population Distribution Dynamics in Genetic Algorithms with High-Probability Mutation
Nicolae-Eugen Croitoru
- 01 Sep 2019
TL;DR: This paper contains an investigation into the GA population dynamics induced by very high mutation operator probabilities (≈ 0.95), used to characterise multiple parameter variants for a Simple Genetic Algorithm, contrasting low-and high-probability mutation, and low- and high-entropy mutation.
3
References
Probability Collectives in Optimization
TL;DR: Some of the work that has been done on Probability Collectives is reviewed, in particular presenting some of the many experiments that have demonstrated its power.
14
The convergence analysis and specification of the Population-Based Incremental Learning algorithm
Helong Li,Sam Kwong,Yi Hong +2 more
TL;DR: This paper presents a meaningful discussion on how to establish a unified convergence theory of PBIL that is not affected by the population and the selected individuals.
7
•Posted Content
Bias-Variance Techniques for Monte Carlo Optimization: Cross-validation for the CE Method
TL;DR: The technique of cross-validation is used, a technique based on the bias-variance tradeoff, to significantly improve the performance of the Cross Entropy (CE) method, which is an MCO algorithm.
5
Combinatorial Optimization via Cross-Entropy
Reuven Y. Rubinstein,Dirk P. Kroese +1 more
- 01 Jan 2004
TL;DR: This chapter shows how the CE method can be easily transformed into an efficient and versatile randomized algorithm for solving optimization problems, in particular combinatorial optimization problems.
Comparing Bayes model averaging and stacking when model approximation error cannot be ignored
TL;DR: Bayes Model Averaging is compared to a non-Bayes form of model averaging called stacking and the results suggest the stacking has better robustness properties than BMA in the most important settings.