Deep Deducing on CartPole, MountainCar, Pendulum and Acrobot

This repository contains codes of deep deducing solving the classic control problems. To reproduce the results in the paper, simply run Deducing.py

Requirements

Packages requirements:

numpy

scipy

Learning phase

To train sets of weight matrices in the paper, run this command:

Learning_xxx.py            

Deducing phase

To use sets of trained weight matrices to start deducing, run this command:

Deducing.py