What You'll Learn

✔ Build a neural network from first principles — perceptrons, logic gates, linear separability, and the foundations of modern deep learning ✔ Implement matrix mathematics in Java — create your own matrix class, transformations, and optimized multiplication ✔ Use activation functions like ReLU and Softmax — implement them from scratch and understand why they matter ✔ Apply information theory and cross‑entropy — entropy, symbol spaces, optimal encoding, and loss functions ✔ Code backpropagation step by step — gradients, chain rule, weighted sums, and full backward passes ✔ Load and process the MNIST dataset — read binary files, parse metadata, batch data, and prepare inputs ✔ Train a network to recognise handwritten digits — run epochs, batches, evaluate loss, and track accuracy ✔ Use multithreading to speed up training — concurrent batch execution and safe locking ✔ Build a reusable neural‑network engine — layers, loaders, configuration, saving/loading models

Who This Course Is For

✔ Java developers who want to understand neural networks at a deep, mathematical level ✔ Programmers moving into machine learning who prefer learning by building instead of relying on libraries ✔ Self‑taught learners who want a clear, structured path through the fundamentals of neural networks ✔ Engineers who want to see how ML works under the hood rather than treating it as a black box ✔ Anyone frustrated by Python‑only ML tutorials and looking for a Java‑based alternative ✔ Developers who enjoy building systems from scratch and want a reusable neural‑network engine they fully understand

The syllabus

11 Chapters · 196 Lessons

  1. 1

    Introduction and Perceptron

    1. (Included in full purchase)

      Introduction

    2. (Included in full purchase)

      Why Write a Neural Network

    3. (Included in full purchase)

      Getting the Most Out of This Course

    4. (Included in full purchase)

      Java vs Python

    5. (Included in full purchase)

      Neurons

    6. (Included in full purchase)

      Perceptron

    7. (Included in full purchase)

      Project With JUnit Support

    8. (Included in full purchase)

      Coding Perceptron

    9. (Included in full purchase)

      Where to Find the Source Code

    10. (Included in full purchase)

      Eclipse Formatters

    11. (Included in full purchase)

      Logic Gates

    12. (Included in full purchase)

      Perceptron AND

    13. (Included in full purchase)

      OR NOR NAND

    14. (Included in full purchase)

      XOR and XNOR

    15. (Included in full purchase)

      Linear Separability

    16. (Included in full purchase)

      Some Layer Terminology

    17. (Included in full purchase)

      Labelling Weights

    18. (Included in full purchase)

      Matrices

    19. (Included in full purchase)

      Some Mathematical Terminology

  2. 2

    Matrix Mathematics

    1. (Included in full purchase)

      A Matrix Class

    2. (Included in full purchase)

      Initialising the Matrix

    3. (Included in full purchase)

      Matrix toString Method

    4. (Included in full purchase)

      Testing the toString Method

    5. (Included in full purchase)

      The Apply Method

    6. (Included in full purchase)

      Multiplying Matrices By a Value

    7. (Included in full purchase)

      Comparing Matrices

    8. (Included in full purchase)

      Using the Equals Method

    9. (Included in full purchase)

      Adding Matrices

    10. (Included in full purchase)

      Motivation for Matrix Multiplication

    11. (Included in full purchase)

      Multiplying the Tables

    12. (Included in full purchase)

      Matrix Multiplication

    13. (Included in full purchase)

      Matrix Multiplication Rule

    14. (Included in full purchase)

      Matrix Multiplication Summary

    15. (Included in full purchase)

      Matrix Multiplication Examples

    16. (Included in full purchase)

      Assertions

    17. (Included in full purchase)

      2D to 1D

    18. (Included in full purchase)

      Iterating Over Multiplicand Rows

    19. (Included in full purchase)

      Completing the Multiplication Implementation

    20. (Included in full purchase)

      Timing Matrix Multiplication

    21. (Included in full purchase)

      Optimising Matrix Multiplication

  3. 3

    Activation Functions

    1. (Included in full purchase)

      Neural Net Test Class

    2. (Included in full purchase)

      Modifying Matrices

    3. (Included in full purchase)

      Adding Bias

    4. (Included in full purchase)

      Multiple Columns of Input

    5. (Included in full purchase)

      ReLu

    6. (Included in full purchase)

      A ReLu Test

    7. (Included in full purchase)

      Matrix forEach

    8. (Included in full purchase)

      Implementing ReLu

    9. (Included in full purchase)

      Introducing Softmax

    10. (Included in full purchase)

      Softmax Worked Example

    11. (Included in full purchase)

      Summing Columns

    12. (Included in full purchase)

      Implementing Softmax

    13. (Included in full purchase)

      Testing Softmax

    14. (Included in full purchase)

      The Engine

    15. (Included in full purchase)

      Deciding Weight Matrix Sizes

    16. (Included in full purchase)

      An Untrained Network

    17. (Included in full purchase)

      Configuring Dense Layers

    18. (Included in full purchase)

      Adding Multiple Layers

    19. (Included in full purchase)

      Running the Engine

  4. 4

    Information Theory and Cross Entropy

    1. (Included in full purchase)

      Mean Squares Loss

    2. (Included in full purchase)

      What is Information

    3. (Included in full purchase)

      Symbol Spaces

    4. (Included in full purchase)

      Entropy

    5. (Included in full purchase)

      An Optimal Encoding Strategy

    6. (Included in full purchase)

      Unequally Probable Symbols

    7. (Included in full purchase)

      Calculating the Information Assoicated With a Symbol

    8. (Included in full purchase)

      Entropy for Unequally Probable Symbols

    9. (Included in full purchase)

      Introducing Cross Entropy

    10. (Included in full purchase)

      Cross Entropy Example

    11. (Included in full purchase)

      Cross Entropy as a Loss Function

    12. (Included in full purchase)

      Implementing Cross Entropy

    13. (Included in full purchase)

      A Cross Entropy Test

    14. (Included in full purchase)

      Implementing the Cross Entropy Test

  5. 5

    Calculus and Backpropagation

    1. (Included in full purchase)

      Training the Network

    2. (Included in full purchase)

      Gradient Descent

    3. (Included in full purchase)

      Gradients and Neural Networks

    4. (Included in full purchase)

      A Calculus Class

    5. (Included in full purchase)

      Implementing Differentiation

    6. (Included in full purchase)

      Basic Mathematical Notation

    7. (Included in full purchase)

      Partial Derivatives

    8. (Included in full purchase)

      Overview of a 3 Layer Network

    9. (Included in full purchase)

      The Network as a Transform

    10. (Included in full purchase)

      Approximator

    11. (Included in full purchase)

      Mock Expected Data

    12. (Included in full purchase)

      Implementing the Transform

    13. (Included in full purchase)

      Examining Loss

    14. (Included in full purchase)

      An AddIncrement Method

    15. (Included in full purchase)

      Completing the Approximator

    16. (Included in full purchase)

      Recap and Natural Logarithms

    17. (Included in full purchase)

      Finishing the Approximator Test

    18. (Included in full purchase)

      Back Propagation

    19. (Included in full purchase)

      Obtaining Softmax Cross Entropy Gradient

    20. (Included in full purchase)

      Backpropagating Errors Through Weighted Sums

    21. (Included in full purchase)

      Introducing the Chain Rule

    22. (Included in full purchase)

      Programming the Chain Rule

    23. (Included in full purchase)

      Chain Rule for Functions of Multiple Variables

    24. (Included in full purchase)

      Programming Multi Variable Chain Rule

    25. (Included in full purchase)

      More Mathematical Terminology

    26. (Included in full purchase)

      Matrix Transpose

    27. (Included in full purchase)

      Implementing Transpose

    28. (Included in full purchase)

      Backpropagation Through Weights

    29. (Included in full purchase)

      Calculating the Backpropagation Transform for Weights

    30. (Included in full purchase)

      Creating a Test for Weights Backprop

    31. (Included in full purchase)

      Creating Weights and Biases

    32. (Included in full purchase)

      Completing the Weights Backprop Test

    33. (Included in full purchase)

      Improving the Weight Backprop Test

    34. (Included in full purchase)

      Backpropagating Through ReLu

    35. (Included in full purchase)

      Implementing Backpropagation Through ReLu

  6. 6

    The Neural Network Engine

    1. (Included in full purchase)

      Adding a RunBackwards Method

    2. (Included in full purchase)

      A Test Data Class

    3. (Included in full purchase)

      A Batch Result Class

    4. (Included in full purchase)

      Configuring the Engine

    5. (Included in full purchase)

      Checking the Loss and Final Transform

    6. (Included in full purchase)

      Storing Errors

    7. (Included in full purchase)

      Initiating Backpropagation

    8. (Included in full purchase)

      Checking Initial Backpropagation

    9. (Included in full purchase)

      Adding Backpropagation Through Weights

    10. (Included in full purchase)

      Adding Backpropagation Through ReLu

    11. (Included in full purchase)

      Turning Off Backpropagation to Input

  7. 7

    Training the Neural Network

    1. (Included in full purchase)

      Calculating the Weight Gradients from the Error

    2. (Included in full purchase)

      Practical Weight Gradient Calculation via Coding

    3. (Included in full purchase)

      Completing the Weight Gradients Test

    4. (Included in full purchase)

      A Useful Fact Involving Matrix Transpose

    5. (Included in full purchase)

      Online Training versus Batch Training

    6. (Included in full purchase)

      Training the Network in Full

    7. (Included in full purchase)

      Testing Average Column

    8. (Included in full purchase)

      Averaging Columns

    9. (Included in full purchase)

      Evaluating Average Loss

    10. (Included in full purchase)

      Getting Weight Inputs and Errors

    11. (Included in full purchase)

      Generating Trainable Data

    12. (Included in full purchase)

      Configuring the Engine Test

    13. (Included in full purchase)

      Finishing the Adjust Method

    14. (Included in full purchase)

      Getting Greatest Row Numbers

    15. (Included in full purchase)

      Adding Percent Correct

    16. (Included in full purchase)

      Repeated Training

    17. (Included in full purchase)

      Generating Spherically Symmetric Distributions

    18. (Included in full purchase)

      Better Training Data

    19. (Included in full purchase)

      Trying the New Data

    20. (Included in full purchase)

      A Running Averages Class

    21. (Included in full purchase)

      Using Running Averages

    22. (Included in full purchase)

      Variable Learning Rates

    23. (Included in full purchase)

      Testing Engine Performance

  8. 8

    Epochs, Batches and Multithreading

    1. (Included in full purchase)

      The MetaData Interface

    2. (Included in full purchase)

      The BatchData Interface

    3. (Included in full purchase)

      The Loader Interface

    4. (Included in full purchase)

      The Neural Network Class

    5. (Included in full purchase)

      Initializing Matrixes with Double Arrays

    6. (Included in full purchase)

      Generating Training Arrays

    7. (Included in full purchase)

      Some Abstract Convenience Classes

    8. (Included in full purchase)

      Creating a Test Loader

    9. (Included in full purchase)

      Completing the Test Loader

    10. (Included in full purchase)

      Testing the Test Loader

    11. (Included in full purchase)

      Summing Matrix Elements

    12. (Included in full purchase)

      Finishing the Loader Test

    13. (Included in full purchase)

      The Fit Method

    14. (Included in full purchase)

      Running with the Test Loader

    15. (Included in full purchase)

      Running the Epochs

    16. (Included in full purchase)

      Running the Batches

    17. (Included in full purchase)

      Implementing RunBatch

    18. (Included in full purchase)

      Adding Multithreading

    19. (Included in full purchase)

      Consuming the Batches

    20. (Included in full purchase)

      Fixing the Execution Exception

    21. (Included in full purchase)

      Outputting Metrics

    22. (Included in full purchase)

      Preventing NaN

    23. (Included in full purchase)

      Improving toString

    24. (Included in full purchase)

      Saving the Network

    25. (Included in full purchase)

      Adding a Load Method

    26. (Included in full purchase)

      Dealing with the Lock

    27. (Included in full purchase)

      Adding a Predict Method

  9. 9

    Loading MNIST Images

    1. (Included in full purchase)

      The MNIST Dataset

    2. (Included in full purchase)

      Command Line Arguments

    3. (Included in full purchase)

      Creating the Image Loader

    4. (Included in full purchase)

      Opening the MNIST files

    5. (Included in full purchase)

      The MNIST File Format

    6. (Included in full purchase)

      Reading Label Metadata

    7. (Included in full purchase)

      Reading the Image MetaData

    8. (Included in full purchase)

      Storing the MetaData

    9. (Included in full purchase)

      Structuring Batch Reading

    10. (Included in full purchase)

      Reading Image Data

    11. (Included in full purchase)

      Converting Bytes to Doubles

    12. (Included in full purchase)

      Reading Batches

    13. (Included in full purchase)

      Reading the Labels

    14. (Included in full purchase)

      An ImageWriter Class

    15. (Included in full purchase)

      Writing Images

    16. (Included in full purchase)

      Determining Canvas Width

    17. (Included in full purchase)

      Calculating Pixel Location

    18. (Included in full purchase)

      Writing the Montage Images

    19. (Included in full purchase)

      Converting One Hot to Int

    20. (Included in full purchase)

      Writing the Labels

  10. 10

    Putting It All Together

    1. (Included in full purchase)

      Putting it All Together

    2. (Included in full purchase)

      Fixing a Nasty Bug

    3. (Included in full purchase)

      Scaling Initial Weights

    4. (Included in full purchase)

      Improving One Hot Conversion

    5. (Included in full purchase)

      Getting Predictions

    6. (Included in full purchase)

      Colorising the Images

  11. 11

    Conclusion

    1. (Included in full purchase)

      Summary

Meet Your Instructor, John Purcell

I’ve taught many thousands of students through programming courses and YouTube tutorials. My focus is always the same: clear explanations, practical examples, and a structured path that makes learning feel natural.

Start Learning Today

Start building your own neural‑network engine in Java and take your skills to the next level with hands‑on, first‑principles learning.

€59,00

Enroll Now