Are you ready to start your machine learning (ML) journey? This course will give you an accessible entry point into this revolutionary field. When you complete this course, you’ll understand core concepts in machine learning and, fundamentally, what are machine learning algorithms trying to do. With a focus on 'math light' explainers, hands-on exercises and real-world examples, our self-paced approach will give you a solid foundation in the core concepts of modern ML.

Machine Learning 101 : Why, What and How?

Section Overview

Supervised Vs Unsupervised Learning - Explainer

K-Nearest Neighbour Explainer

K-Nearest Neighbour Walkthrough

K-Means Explainer

Trying out Clustering with K-Means

Data Preparation, Representation and Model Selection

Loss Functions

Training, Optimisation & Hyper-parameter Tuning

Exploring If Models are Any Good - Training Curves

Is My Model any Good - Validation Plots

Conclusion, Certificate, and What Next?

digiLab

Learn key machine learning concepts and build a strong foundation for future learning.

Getting Started in Machine Learning

Linear models might not seem that exciting, but they provide the foundation you need to understand advanced topics like Deep Neural Networks or Gaussian Processes further down the line. In this course, we’ll work through key concepts and techniques, including Predictive Power Score (or PPS) and Principle Component Analysis (PCA). We'll also cover the important topic of over-fitting and how we address this using regularisation methods.

The Foundations of ML - Curve Fitting

Regression Walkthrough

Underfitting and Overfitting

Controlling ML Models - Regularisation in Practise

Exploring Simple Relationships - Correlations

Finding Nonlinear Relations - Predictive Power Score

CSV file with the data used in this lesson

Correlation & PPS Walkthrough

From Regression to Classification - Logistic Regression

Logistic Regression in Wild

Looking through the Right Lens - Principle Component Analysis

Looking in the Right Direction - Exploring PCA

Understanding linear models is the key to unlocking the full power of machine learning.

Using General Linear Models for Machine Learning

With a parametric model, we guess the shape of the function and fit with suitable parameters. With a non-parametric model, we outsource the guesswork to the model itself: the machine does the thinking! As a first foray into non-parametric algorithms, we look at the kernel trick, which amplifies our input data by combining all the features under the hood. The first model we’ll examine is Support Vector Machines for classification and regression tasks.

Don't Be Fooled by the Kernel Trick

A mindmap of what we covered in this lesson

Projecting data features into higher dimensions

Introducing: The Kernel Trick

Choosing Between Kernel Functions

A Code Walkthrough for Kernel Ridge Regression

Applying the Kernel Trick in the Wild

Introducing: Support Vector Machines

Support Vector Machines to Maximise Decision Margins

A Code Walkthrough for SVMs

Overlapping Classes and Kernel SVMs

Experimenting with Overlapping Class Distributions

Using Kernel SVMs for Non-Linear Predictions

Support Vector Machines in the Wild

Solving Regression Problems with SVMs

Comparing Least-Squares with SVM Regression

Outsourcing the Guesswork: A Dive into the Kernel Trick and Support Vector Machines.

The Kernel Trick: A first look at flexible machine learning

The archetypal and top-performing machine learning algorithms are based on the humble decision tree. Trees are explainable, intuitive algorithms which are cheap to train and run. But on their own they need support. In this course, we’ll investigate the fundamentals of Decision Trees and then explore their extension to both Random Forests and Gradient Boosted Trees (inc. ‘XGBoost’). These two algorithms are the gold medalists of ML models on any leaderboard, an essential tool for your collection.

Introducing Decision Trees

Decision Trees in Everyday Thinking

Machine-Designed Decision Trees

Classification Problems with Decision Trees: A Code Walkthrough

Regression Problems with Decision Trees: A Code Walkthrough

Ensemble Methods: Machine Learning and Democracy

Random Forests: Decisions Don't Fall Far from the Decision Tree

Random Forests out in the Wild

Interpolation Through a Random Forest

Give Yourself a Gradient Boost

Auto-Correction in a Forest of Stumps

Gradient Boosting by Hand: Code Example

XGBoost in the Wild

Cross validate with the XGBoost API

Gold Standard Algorithms: Harnessing the Power of Random Forests and Gradient-Boosted Models.

Tree-based algorithms: Random Forests and Gradient Boosted Trees

If you want to deploy AI in the Wild with confidence, then you need Uncertainty Quantification (UQ). Fundamentally, we must be able to assess the reliability of our estimations. This course will introduce you to the many faces of Bayesian Inference - the gold standard of robust and explainable UQ - from the theoretical foundations to Bayesian parameter estimation.

Introduction to Uncertainty Quantification

The Importance of Being Bayes: Introducing Bayes' Theorem

Unpacking Bayes' Theorem: Prior, Likelihood and Posterior

Prior and Posterior - Coin Flipping Walkthrough

Linear Least Squares Breakdown and Bayesian Linear Regression

Bayesian Linear Regression: A Coding Walkthrough

Bayesian Parameter Estimation

Bayesian Parameter Estimation: Walkthrough with COVID Example

If you want to deploy AI with confidence, then you need Uncertainty Quantification (UQ).

An Introduction to Bayesian Uncertainty Quantification

In this course, we venture into the realm of AI Ethics, examining and attempting to unpack the societal implications surrounding data and privacy in our increasingly digital lives. We’ll probe the intriguing intersection of AI and creativity: can machine learning models genuinely exhibit creativity or are we attributing human characteristics to algorithms? Hear firsthand from artists about the influence of generative AI on their work. With a blend of theoretical discussions and practical insights, this course provides food for thought on the ethical dimensions of modern machine learning.

The interplay between ML, AI, and AGI

Creativity of, around, and with AI

Data and privacy concerns for ML tools

Exploring Boundaries and Unraveling the Ethical Dimensions of AI and Machine Learning.

AI Ethics: Fairness and Balance in Modern Machine Learning

This course will guide you through the complexities of building and deploying machine learning models to production. We’ll use TensorFlow, an industry-standard library, to build our neural network. We’ll explore the intricacies of dataset compilation, teaching you the ins and outs of the Bing API for creating your own custom datasets. We’ll also discuss how to reduce the footprint of your ML models using ONNX, as well as saving and retrieving them from a data store using the S3 buckets API. When you complete this course, you’ll have the tools to develop your own cloud-based, machine learning-powered web applications.

Project setup

Training a convolutional neural network

Classifying an image

Visualising a model

Continuing training

Adding a React user interface

Serving our application

Calling the backend from the frontend

Accepting an image

Uploading an image with the UI

Deploying as a serverless image

Returning the class predictions

Returning internal model activations

Displaying the model’s activation activity

Code refactor

Code optimisation

Build, Train and Deploy a Convolutional Neural Network Image Classifier.

Using TensorFlow to Build a Production-Ready Neural Network-Powered Web App

The foundation of every successful machine learning model is clean, high-quality data. But what does clean data look like, and how can you ensure the data you're working with meets that standard? Whether you're dealing with tabular data in Excel or tackling larger datasets in Python, understanding effective strategies for data cleaning is crucial. In this course, we’ll equip you with the essential techniques to turn messy, incomplete, and inconsistent data into a valuable resource.

What Does Clean Data Look Like in Microsoft Excel?

Supercharge Your Microsoft Excel Data Cleaning with Python

How to Clean Data in Excel Using Python: The Basics

Extracting Values From Free Text: Introduction to REGEX

Manipulating and working with Pandas dataframes

Managing Missing Values Part 1

Managing Missing Values Part 2

Obtaining Data Part 1: Working with Files

Obtaining Data Part 2: Working with Web Data

Case Study: Bringing it All Together

Master essential data cleaning techniques and turn messy data into a valuable resource with effective cleaning techniques in Excel and Python.

Strategies and Techniques for Data Cleaning in Excel and Python

The aim of this course is not to teach you data science or machine learning but rather to lay the groundwork allowing you to work efficiently through the following courses with the rest of the digiLab team. We’ll start with the basics and cover getting Python installed and your coding environment set up. Then, we take a whistle-stop tour through Python fundamentals to familiarise you with basic programming concepts like loops, branching logic, functions and even some object-oriented programming.

The purpose of this course

Our approach to learning Python

General tips for success

Use what works for you...but we have suggestions

Additional resources and getting help

Environment setup - section overview

Download the Anaconda Python distribution

Downloading Python and Visual Studio Code

Python environments

Welcome to your new home - VS Code

Getting to know Jupyter Notebooks

Python basics - section overview

Variables and Python data types

Functions in Python

Loops in Python

Control flow

Solution walkthrough

Basic object-oriented programming concepts

Get your coding environment set up and learn the basics of Python.

Pre-flight checks - Getting Started & Python Basics

Deep neural networks are now solving problems that, until recently, were out of reach. You’ve no doubt seen them pop up in the news as part of the AI revolution. But let’s take a step back. What are neural networks? What is a neurone? What makes them deep? And how can we train and use these models? In this course, we start at the beginning, explaining the fundamental concepts and introducing programming essentials to prepare you to deploy and understand deep neural networks.

Course Introduction

Learn the fundamentals of deep neural networks, including selecting architecture and training algorithms.

Deep Learning 101: Getting Started with Neural Networks

Delve into the fascinating realm of Reinforcement Learning (RL). You'll gain insights into the processes that enable machines to learn from their environment, optimizing their actions for maximum reward. This course covers essential RL strategies, from understanding optimization and reward shaping to policy-based and value-based learning. By the end, you'll be equipped with practical RL skills, having built an agent that can win the game of Pacman using OpenAI Gym and RLlib. This course will set you on a confident path to leveraging reinforcement learning in your future AI projects.

Adapting, Evolving, Winning: An Introduction to Reinforcement Learning Techniques

Reinforcement Learning 101: Learn RL by building an AI agent to win at Pacman.

One of the biggest obstacles to starting your exciting journey in machine learning is confidence in math. At digiLab, we believe that you don't need to know extensive math to get started. In this course, we provide you with the essential tools and concepts you need to begin. We break down our learning into small, bite-sized pieces, showing you that Linear Algebra, Calculus, and Probability won't stand in your way. This material will be released gradually over time. Also, watch out for the math masterclasses conducted by the digiLab team, which will be held live.

Cover all of the basics you missed at school, or at least enough to manage machine learning.

Just Enough Maths for Machine Learning

In this course, we will teach you about the fascinating world of Gaussian Process Regression and Classification, a nonparametric Machine Learning technique that elegantly addresses predictive uncertainty with mathematical rigor. This powerful method can reveal unique insights into complex phenomena. We will guide you through Gaussian Process design, hyperparameter optimization, model selection, and diagnostics, empowering you to harness this versatile technique with confidence.

Master Gaussian Process Regression and Classification and learn how to design and diagnose your models with confidence.

Gaussian Processes - Freeing your Functions

Large Language Models (LLMs) have demonstrated remarkable human-like capabilities in answering a variety of questions and fulfilling requests. However, defining what constitutes a “good” answer can be subjective. In this course, we'll explore how Reinforcement Learning with Human Feedback has enhanced the capabilities of these models. We'll also delve into how this provides a framework for fine-tuning LLMs for specialist tasks or for integrating additional data in specific contexts. Throughout the course, you will work alongside me to build the digiLab Academy chatbot. By the end, you'll be able to write the code for your own applications and have a deep understanding of the fascinating mechanics behind these methods.

Optimising AI with Human Feedback: Specialise your LLM on your data or for your task!

Reinforcement Learning with Human Feedback: Tuning a Custom Large Language Model

twinLab is digiLab’s state-of-the-art machine learning library. Developed in-house by our data science team, twinLab is the toolset we use to solve our clients’ most challenging problems. Built to tackle real-world, data science ‘in the wild’ problems, from aerospace to fusion energy, twinLab is what we use to decode the data and unlock value for our customers. Now we want to give you access and teach you how to use it! The aim of this short course is to get you confidently applying twinLab in your own data science pipeline. 

Mastering Digital Twinning: Real-World Applications with twinLab.

twinLab - Welcome to Machine Learning

This beginner-friendly course demystifies key machine learning (ML) concepts, introducing you to Supervised and Unsupervised Learning, Regression and Classification, and efficient Machine Learning Workflows. Gain a strong foundation in Generalized Linear Models, Introductory Statistics, Exploring Correlations, and Predictive Power Scores. As your first foray into ML, this course will expertly guide you through the essentials.

From Zero to Sixty - Foundations in Machine Learning

In our first exploration of advanced machine learning (ML), we focus on non-parametric modelling, allowing machines to choose models themselves. Some of the most powerful practical ML tools include kernel methods, support vector machines, decision trees, forests, and gradient boosting ensemble methods. This course covers the core concepts and provides hands-on coding examples for applying these techniques.

Introducing: Decision Trees

Learn how to apply advanced machine learning models to real-world problems.

Machine Thinking - Machine Learning Models for Professionals

This course introduces you to the exciting world of Gaussian Processes, providing a practical introduction to the world of Uncertainty Quantification. We’ll discuss the fundamentals of Gaussian Distributions and explore the intricacies of multivariate correlations and conditioning. As we navigate through higher dimensions and delve into practical examples, you'll gain confidence in handling uncertainty in Machine Learning. Whether you're a beginner or an advanced learner, this course offers a blend of theory and practice, guiding you to master Gaussian Processes and embrace uncertainty with confidence.

Decoding Gaussian Processes: Embrace Uncertainty with Confidence.

Getting Started with Gaussian Processes and Uncertainty Quantification

Getting Started in Machine Learning

Learn key machine learning concepts and build a strong foundation for future learning.

AI in the Wild

Get Started for Free