Genetic algorithm to optimize hyperparameters in a Deep learning net

Hi all

I have made a network using the deep learning toolbox with various hyperparameters such as mini-batch size and number of neurons per layer etc. Currently I am using a grid search to find the best combinations but is there a way to use the genetic algortihm in the Optimization toolbox to find the best architecture? I have read similar threads but none seem to explain how one could achieve this. Any explanations or examples of code would be really appreciated!

[Muhammad Mohsin Khan]

Hello Raidoo,

Will you like to share your Grid search or GA code for deep learning time series data ? If can not any other help ?

Thanks

Answer 1

[Dinesh Yadav]

Hi Renita,

Kindly go through the following code in File Exchange

https://www.mathworks.com/matlabcentral/fileexchange/72283-hyperparameter-optimization-for-single-layer-neural-networks/?s_tid=LandingPageTabfx

And also this thread

https://www.mathworks.com/matlabcentral/answers/304095-i-want-to-optimize-a-neural-network-parameters-by-genetic-algorithm-and-i-get-below-code-from-matlab

Hope it gives you some insights.

[Renita Raidoo]

Thank you much! This is very helpful.