Project Setup.

Opening Command Prompt.

Activating Virtual Environment 1.

Installing PyTorch.

Test if ML Agents is properly installed.

Why Use An Environment?

Adding Colliders.

Agent Script Setup.

What are Actions (Unity Learn).

What Are Observations?

SerializedField?

Better Way To Check Collision.

OnEpisodeBegin.

Decision Requester.

Heuristic (1).

Heuristic Test 2.

Heuristic Test 3.

Training Agent 1.

Multiple Environment.

I made a mistake.

Getting Results Form Training .

How to Better Train AI.

Training Agent 3 (1).

How To See How Good The AI Did.

Finial Result (Full Screen).

Intro

Unity Setup

Python

Create Virtual Environment

Install Dependencies

Unity Environment

OnActionReceived

CollectObservations

Create Target

OnTriggerEnter

OnEpisodeBegin

Changes to Agent

Heuristic

Bug Fix

Real Time Training Session

Sped Up Training

Improve Training Time

Sped Up Training

Fixing Reward Issue

Proper Training Run

Using Neural Network

Start Up a New Unity Project

The First Scenario

Run with Ml Agents

Create an Agent Script

Bound Floor

Move Target Agent Script

Sample Neural Network

Anaconda Navigator

Ray Perception Sensor Sensor

Example Scene

Randomization

Making changes to the environment

Prototype using free or inexpensive assets

Start Simple

Multiple Cameras

Mini Map

Design in flexibility in perceptions, actions and rewards

Action parameters

Conclusion

Intro

Setting Scripting Runtime Version

Truffle Prefab

Rotation

Creating the PigArea Script

Configuring Agents

Agent Actions

Difficulty Levels

Training the Model

Training Troubleshooting

Intro

Why use agents in games?

Reinforcement learning with a reward-based signal

Example projects

Curiosity-driven exploration

What is Machine Learning Inference?

Cross-Platform Inference

How do you define curiosity in terms of intrinsic reward?

How deterministic is the behavior result?

Closing remarks

Introduction

Motivation - Why an introduction to ML-Agents?

Artificial Neural Networks

Deep Learning

Types of Deep Learning

Deep Reinforcement Learning

Deep Reinforcement Learning Flowchart

Closed-Loop Decision Diagram

Deep Reinforcement Learning Problem Structure

Deep Reinforcement Learning Sample Problem

To Be Continued...

Intro

ML Agents

Install Git

Clone ML Agents

Create Anaconda Env

Install PyTorch

Install Cuda 11.8

Install ML Agents

Install Protobuf 3.20.3

Open Unity Project

Training PushBlock

Install onnx

Re-run Training

Test Neural Network

Intro

Push Block Example

Demo Overview

Creating Demos

Move Agent

Demonstration Recorder

Demonstration Recording

Demo size comparison

Analyzing Data

Conclusion

Introduction

Download and Install Anaconda

Download ML-Agents Release 19 via GitHub

Setup Anaconda Environment

Create New Unity Project

Install ML-Agents in Unity

Import ML-Agents Examples

Test ML-Agents with Push Block environment

Start Training your own Neural Network

Test your Neural Network in Inference Mode

Intro

Unity ML-Agents Forum

Release 10 Tags

Installation

Opening the Example Project

Wall Jump Demo

The Agent

Jump Functionality

Keyboard Control Demo

Installing ML-Agents

Intro

Installation

Downloading Unity Project

ML-Agents folders

3D Ball example

Pre-trained neural network weights

Activate Conda Environment

Troubleshooting

Checking the saved model

Outro