Hello!
I am using Artificial Neural Networks to predict the hedonic response of images. After I train my model and use the predictions as the fitness criteria to navigate a solution space using Galapagos, after some solutions I get the same prediction value time after time, doesn’t matter the inputs.
It has been happening in Crow, in LunchBoxML, and in a custom C# component, so I guess it is a Grasshopper/Galapagos problem.
Have you experienced something like that?
Thank you!