Skip to content

Latest commit

 

History

History
9 lines (5 loc) · 847 Bytes

README.md

File metadata and controls

9 lines (5 loc) · 847 Bytes
Raw

Reinforcement_Learning_with_Backpropagation

This project consists of three milestone projects

  • Milestone 1: Training a two-layer neural network based on XOR gate truth table with the error-backpropagation algorithm. All the relevant codes and files are in the Backpropagation_Learning folder

  • Milestone 2: Implementeing reinforcement learning for a robot tank in Robocode environment with a look-up table and comparing the reinforcement learning performance between on-policy and off-policy. All the relevant codes and files are in the Reinforcement_Learning_with_LUT folder

  • Milestone 3: Replacing the look-up table with the neural network, which has the optimum hyperparameters, to train the robot tank in the Robocode environment. All the relevant codes and files are in the Reinforcement_Learning_with_Backpropagation folder