The FiniteDiffRegressor – implemented in Python package tisthemachinelearnerblends a finite difference-based training algorithm for (quasi-)randomized artificial neural network models’ weights with any supervised Machine Learning regression model.

More details about the implementation can be found in this technical note:

https://www.researchgate.net/publication/392923564_Backpropagating_quasi-randomized_neural_networks

On GitHub:

https://github.com/Techtonique/tisthemachinelearner

And here is a link to the notebook containing examples of use of FiniteDiffRegressor:

Open In Colab

image-title-here

Comments powered by Talkyard.