Abstract | ||
---|---|---|
Artificial neural networks typically use backpropagation methods for the optimization of weights. In this paper, we aim at investigating the potential of applying the so-called evolutionary strategies (ESs) on the weight optimization task. Three commonly used ESs are tested on a multilayer feedforward network, trained on the well-known MNIST data set. The performance is compared to the Adam algorithm, in which the result shows that although the (1 + 1)-ES exhibits a higher convergence rate in the early stage of the training, it quickly gets stagnated and thus Adam still outperforms ESs at the final stage of the training.
|
Year | DOI | Venue |
---|---|---|
2019 | 10.1145/3319619.3322014 | GECCO |
Keywords | Field | DocType |
Evolution strategy, backpropagation, neural network training | MNIST database,Computer science,Evolution strategy,Rate of convergence,Artificial intelligence,Backpropagation,Artificial neural network,Machine learning,Feed forward | Conference |
ISBN | Citations | PageRank |
978-1-4503-6748-6 | 0 | 0.34 |
References | Authors | |
2 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Hao Wang | 1 | 27 | 10.69 |
Thomas Bäck | 2 | 629 | 86.94 |
Aske Plaat | 3 | 524 | 72.18 |
Michael T. Emmerich | 4 | 21 | 5.20 |
Preuss Mike | 5 | 933 | 81.70 |