All geek questions in one place

Design for teaching the art of artificial intelligence

This is a community wiki whose goal is to create a good design for a machine learning / artificial intelligence system (ML / AI structure).

Please make your contribution to the development of a linguistic agnostic structure that would allow combining several ML / AI algorithms into a single structure that:

  • runs algorithms with a user-defined data set.
  • facilitates training, qualifications and classification.
  • allows users to easily connect new algorithms.
  • can combine or create an ensemble of existing algorithms.
  • can save / load the course of the algorithm (i.e. save the network and the weights of the neural network, save the tree of the decision tree, etc.).

What is a good design for this kind of ML / AI framework?

+4
source share
2 answers

Perhaps you can start by looking at the structure of existing open source ML / AI modules. To name a few: Weka, RapidMiner, KNIME, Orange, ..

+3
source

Here is the one I made for PHP: http://neuralmesh.com

+1
source

Source: https://habr.com/ru/post/1303879/

More articles:


All Articles