AI Brain Project

This project implements a Deep Q-Learning (DQN) based AI brain using PyTorch. It's designed to learn and make decisions in an environment through reinforcement learning.

Features

• Neural Network architecture for Q-value approximation
• Experience Replay for efficient learning
• Deep Q-Learning algorithm implementation
• Action selection using softmax distribution
• Model saving and loading capabilities

Dependencies

• Python 3.x
• PyTorch
• NumPy
• torch
• kivy

Installation

1. Ensure you have Python 3.x installed.
2. Install the required packages:

   pip install numpy torch kivy
   

Usage

1. Import the Dqn class from the main script:

   from main import Dqn

2. Initialize the DQN agent:

   input_size = <your_input_size>
   nb_action = <number_of_possible_actions>
   gamma = <discount_factor>
   
   agent = Dqn(input_size, nb_action, gamma)

3. Use the agent in your environment:

   state = <current_state>
   action = agent.select_action(state)
   
   # After taking the action and observing the result:
   new_state = <new_state>
   reward = <observed_reward>
   
   next_action = agent.update(reward, new_state)

4. Save and load the model:

   # To save
   agent.save()
   
   # To load
   agent.load()

Components

Network

The Network class defines the neural network architecture used for Q-value approximation. It consists of two fully connected layers.

ReplayMemory

The ReplayMemory class implements experience replay, storing and sampling past experiences for batch learning.

Dqn

The Dqn class ties everything together, implementing the Deep Q-Learning algorithm. It handles action selection, learning, and model updates.

Customization

You can customize various aspects of the AI brain:

• Adjust the network architecture in the Network class
• Modify the learning rate in the Dqn class initialization
• Change the temperature parameter in the select_action method
• Adjust the replay memory size and batch size in the Dqn class