Skip to content
Concept-Lab
← Back to Machine Learning

Machine Learning

Supervised learning, linear and logistic regression, gradient descent, cost functions, regularisation, and the full breadth of Andrew Ng's ML Specialisation.

0 / 114 completed0%
114 remainingShowing 114 of 114 nodes

Concepts Covered

Supervised Learning Algorithms

0/1 done
1

Introduction to Machine Learning

What ML is, where you already use it daily, and why this matters.

Interactive

Advanced Learning Algorithms

0/1 done
1

Introduction to Advanced Learning Algorithms

Course 2 overview: neural networks, decision trees, and practical ML system advice.

Interactive

Unsupervised, Recommenders & Reinforcement

0/1 done
1

Welcome

Course map: unsupervised learning first, recommender systems next, reinforcement learning after that.

Theory

Supervised Learning Algorithms

0/1 done
2

Why Machine Learning Matters

ML as the dominant path to AI; the $13-trillion opportunity ahead.

Interactive

Advanced Learning Algorithms

0/1 done
2

Neurons and the Brain

The biological motivation behind neural networks and why deep learning took off when it did.

Theory

Unsupervised, Recommenders & Reinforcement

0/1 done
2

What Is Clustering?

Clustering finds structure in unlabeled data by grouping similar points together.

Lab

Supervised Learning Algorithms

0/1 done
3

ML Definition & Types

Supervised, unsupervised, and reinforcement learning — when to use each.

Lab

Advanced Learning Algorithms

0/1 done
3

Neural Networks: Demand Prediction

Building intuition for neural networks using a T-shirt top-seller prediction example.

Interactive

Unsupervised, Recommenders & Reinforcement

0/1 done
3

K-Means Intuition

K-means alternates between assigning points to nearest centroids and moving centroids to cluster means.

Lab

Supervised Learning Algorithms

0/1 done
4

Supervised Learning — Regression

Predicting continuous output values — the engine behind 99% of ML's economic value.

Interactive

Advanced Learning Algorithms

0/1 done
4

Recognising Images with Neural Networks

How neural networks build up visual understanding layer by layer — edges, parts, then faces.

Theory

Unsupervised, Recommenders & Reinforcement

0/1 done
4

K-Means Algorithm

Formal K-means procedure with assignment equations, centroid updates, and empty-cluster handling.

Lab

Supervised Learning Algorithms

0/1 done
5

Supervised Learning — Classification

Predicting discrete categories rather than continuous values.

Lab

Advanced Learning Algorithms

0/1 done
5

Neural Network Layers

Layer notation, superscripts, and how a single hidden layer computes its activations.

Interactive

Unsupervised, Recommenders & Reinforcement

0/1 done
5

Optimization Objective

K-means minimizes distortion: average squared distance from each point to its assigned centroid.

Lab

Supervised Learning Algorithms

0/1 done
6

Unsupervised Learning

Finding hidden structure in data with no labels — clustering, anomaly detection, and more.

Interactive

Advanced Learning Algorithms

0/1 done
6

More Complex Neural Networks

Multi-layer networks, counting conventions, and the general activation formula for any layer.

Theory

Unsupervised, Recommenders & Reinforcement

0/1 done
6

Initializing K-Means

Initialization quality strongly affects final clustering; multi-start runs improve robustness.

Lab

Supervised Learning Algorithms

0/1 done
7

Unsupervised — Anomaly Detection

Detecting fraud, defects, and outliers — the three types of unsupervised learning.

Interactive

Advanced Learning Algorithms

0/1 done
7

Inference & Forward Propagation

The algorithm for making predictions: computing activations left to right through all layers.

Interactive

Unsupervised, Recommenders & Reinforcement

0/1 done
7

Choosing the Number of Clusters

Choosing K is often ambiguous; combine elbow hints with downstream business tradeoffs.

Lab

Supervised Learning Algorithms

0/1 done
8

Jupyter Labs & Dev Environment

The industry-standard ML environment — the exact same tool used at Google, Meta, and Amazon.

Interactive

Advanced Learning Algorithms

0/1 done
8

Inference in Code (TensorFlow)

Implementing forward propagation in TensorFlow with Dense layers — the coffee roasting example.

Theory

Unsupervised, Recommenders & Reinforcement

0/1 done
8

Finding Unusual Events

Anomaly detection learns normal behavior and flags low-probability events for inspection.

Lab

Supervised Learning Algorithms

0/1 done
9

Linear Regression Pipeline

Your first supervised learning model — probably the most widely used ML algorithm in the world.

Interactive

Advanced Learning Algorithms

0/1 done
9

Data Representation in TensorFlow

NumPy 1D vectors vs 2D matrices, TensorFlow tensors, and why the double bracket matters.

Theory

Unsupervised, Recommenders & Reinforcement

0/1 done
9

Gaussian (Normal) Distribution

Gaussian distributions model feature likelihood via mean and variance, forming the basis of simple anomaly scoring.

Lab

Supervised Learning Algorithms

0/1 done
10

The Supervised Learning Pipeline

How supervised learning actually works end-to-end — training set in, function out.

Interactive

Advanced Learning Algorithms

0/1 done
10

Building a Neural Network in TensorFlow

The Sequential API: string layers together, compile, fit, and predict in one workflow.

Interactive

Unsupervised, Recommenders & Reinforcement

0/1 done
10

Anomaly Detection Algorithm

Fit one Gaussian per feature, multiply densities into p(x), then classify with epsilon threshold.

Lab

Supervised Learning Algorithms

0/1 done
11

Cost Function

Measuring how wrong your model is — Mean Squared Error (MSE) explained.

Interactive

Advanced Learning Algorithms

0/1 done
11

Forward Prop: Single Layer from Scratch

Implementing forward propagation in raw Python/NumPy — understanding what TensorFlow does under the hood.

Interactive

Unsupervised, Recommenders & Reinforcement

0/1 done
11

Developing and Evaluating an Anomaly Detection System

Use cross-validation anomalies to tune epsilon and features; evaluate with skew-aware metrics like precision, recall, and F1.

Lab

Supervised Learning Algorithms

0/1 done
12

Cost Function Intuition

What the cost function looks like — and why the bowl shape matters.

Interactive

Advanced Learning Algorithms

0/1 done
12

General Forward Propagation

The dense() function: a reusable loop-based implementation of any layer using a weight matrix.

Interactive

Unsupervised, Recommenders & Reinforcement

0/1 done
12

Anomaly Detection vs Supervised Learning

Pick anomaly detection for rare and evolving positives; pick supervised learning when positives are sufficiently labeled and stable.

Lab

Supervised Learning Algorithms

0/1 done
13

Cost Visualisation in 3D

Contour plots and the 3D bowl — seeing the optimisation landscape with two parameters.

Interactive

Advanced Learning Algorithms

0/1 done
13

Is There a Path to AGI?

ANI vs AGI, the one learning algorithm hypothesis, and keeping hype calibrated.

Theory

Unsupervised, Recommenders & Reinforcement

0/1 done
13

Choosing What Features to Use

Feature shaping and engineering are critical in anomaly detection; transform skewed variables and iterate via error analysis.

Lab

Supervised Learning Algorithms

0/1 done
14

Parameters, Model & Cost — Together

Connecting the model line, cost function, and contour plot into one unified picture.

Interactive

Advanced Learning Algorithms

0/1 done
14

Vectorised Neural Network Implementation

Why matrix multiplication makes neural networks fast and how NumPy's matmul replaces for-loops.

Theory

Supervised Learning Algorithms

0/1 done
15

Gradient Descent — Concept

The core optimisation algorithm that trains virtually every ML model.

Interactive

Advanced Learning Algorithms

0/1 done
15

Matrix Multiplication

Dot products, transposes, and vector-matrix products — the mathematical building blocks of neural networks.

Theory

Supervised Learning Algorithms

0/1 done
16

Gradient Descent — Update Rule

The actual update equations — the math behind every gradient step.

Lab

Advanced Learning Algorithms

0/1 done
16

Matrix Multiplication Rules

The general formula for AᵀW — computing every element systematically from rows and columns.

Theory

Supervised Learning Algorithms

0/1 done
17

Derivative Intuition for Gradient Descent

The tangent line trick — why the sign and magnitude of the gradient guide every step.

Interactive

Advanced Learning Algorithms

0/1 done
17

Matrix Multiplication in Code

NumPy matmul, the @ operator, and the complete vectorised forward pass in TensorFlow.

Theory

Supervised Learning Algorithms

0/1 done
18

Learning Rate

The most critical hyperparameter — too large diverges, too small barely moves.

Interactive

Advanced Learning Algorithms

0/1 done
18

Training a Neural Network: Overview

The three-step training loop in TensorFlow: specify architecture, compile with loss, fit to data.

Lab

Supervised Learning Algorithms

0/1 done
19

Completing Linear Regression

The complete training loop: model + cost + gradient derivation all in one.

Interactive

Advanced Learning Algorithms

0/1 done
19

Training Details: Loss, Cost, and Backprop

Binary cross-entropy loss, cost function over all examples, and how TensorFlow uses backprop internally.

Lab

Supervised Learning Algorithms

0/1 done
20

Gradient Descent — Live Demo

Watching the algorithm actually run — the parameter trajectory toward the minimum.

Interactive

Advanced Learning Algorithms

0/1 done
20

Alternatives to the Sigmoid Activation

ReLU, linear activation, and why hidden layers should almost never use sigmoid.

Lab

Supervised Learning Algorithms

0/1 done
21

Multiple Linear Regression

Extending to many features simultaneously — the vectorised dot product form.

Lab

Advanced Learning Algorithms

0/1 done
21

Choosing Activation Functions

How to pick the right activation function for output and hidden layers based on what you're predicting.

Lab

Supervised Learning Algorithms

0/1 done
22

Vectorisation

Why vectorised code is 100× faster — numpy and hardware parallelism.

Lab

Advanced Learning Algorithms

0/1 done
22

Why Do We Need Activation Functions?

Without nonlinear activations, a deep neural network collapses into a simple linear model.

Interactive

Supervised Learning Algorithms

0/1 done
23

Vectorisation — Under the Hood

How NumPy, BLAS, and GPU kernels actually execute computations in parallel.

Interactive

Advanced Learning Algorithms

0/1 done
23

Multiclass Classification

When y can take more than two values — classifying digits 0–9, diseases, or defect types.

Interactive

Supervised Learning Algorithms

0/1 done
24

Feature Scaling

Normalising features so gradient descent converges faster — a must-do step.

Interactive

Advanced Learning Algorithms

0/1 done
24

Softmax Regression

The generalization of logistic regression to n classes — computing mutually exclusive class probabilities.

Interactive

Supervised Learning Algorithms

0/1 done
25

Implementing Feature Scaling

Coding z-score normalisation from scratch; using sklearn's StandardScaler.

Interactive

Advanced Learning Algorithms

0/1 done
25

Neural Network with Softmax Output

Plugging softmax into the output layer to build a multiclass neural network.

Interactive

Supervised Learning Algorithms

0/1 done
26

Gradient Descent Convergence

The learning curve — how to tell when training is done and when it's broken.

Interactive

Advanced Learning Algorithms

0/1 done
26

Numerically Stable Softmax

Using from_logits=True to avoid floating-point roundoff errors in softmax and logistic loss.

Interactive

Supervised Learning Algorithms

0/1 done
27

Choosing the Learning Rate

The log-scale sweep strategy for finding a good α systematically.

Interactive

Advanced Learning Algorithms

0/1 done
27

Multi-Label Classification

When one input can have multiple independent labels simultaneously — cars AND pedestrians in one image.

Interactive

Supervised Learning Algorithms

0/1 done
28

Feature Engineering

Creating better input features using domain knowledge — often the biggest performance lever.

Interactive

Advanced Learning Algorithms

0/1 done
28

Adam Optimizer

Adaptive Moment Estimation — the de facto standard optimizer that auto-adjusts per-parameter learning rates.

Interactive

Supervised Learning Algorithms

0/1 done
29

Polynomial Regression

Fitting curves not just lines — by engineering x², x³ as new features.

Interactive

Advanced Learning Algorithms

0/1 done
29

Convolutional Layers

Beyond dense layers — how convolutional layers let neurons see only local regions for speed and robustness.

Interactive

Supervised Learning Algorithms

0/1 done
30

Classification — Deep Dive

Why linear regression fails for classification and what to use instead.

Interactive

Advanced Learning Algorithms

0/1 done
30

What is a Derivative?

Intuitive foundations of calculus derivatives — the engine behind gradient descent and backpropagation.

Interactive

Supervised Learning Algorithms

0/1 done
31

Logistic Regression

The sigmoid function — squashing any real number into a probability [0, 1].

Lab

Advanced Learning Algorithms

0/1 done
31

Computation Graph and Backprop

How neural network frameworks compute gradients efficiently using forward and backward passes through a graph.

Lab

Supervised Learning Algorithms

0/1 done
32

Decision Boundary

Where the model draws the line between classes — linear and non-linear boundaries.

Interactive

Advanced Learning Algorithms

0/1 done
32

Backprop in a Larger Network

Tracing backpropagation through a two-layer network — seeing how gradients flow back to every parameter.

Interactive

Supervised Learning Algorithms

0/1 done
33

Logistic Regression — Cost Function

Why MSE creates non-convex surfaces for classification; introducing log loss.

Lab

Advanced Learning Algorithms

0/1 done
33

Deciding What to Try Next

The systematic approach to ML debugging — why intuition fails and diagnostics save months of wasted effort.

Interactive

Supervised Learning Algorithms

0/1 done
34

Simplified Logistic Loss

Combining the y=0 and y=1 cases into one elegant unified formula.

Lab

Advanced Learning Algorithms

0/1 done
34

Evaluating a Model

Train/test splits and why J_train alone deceives you — measuring generalization systematically.

Interactive

Supervised Learning Algorithms

0/1 done
35

Gradient Descent for Logistic Regression

Same update rule as linear regression — but with sigmoid applied underneath.

Lab

Advanced Learning Algorithms

0/1 done
35

Model Selection and Cross-Validation

The three-way split — why you need a cross-validation set to choose models without contaminating the test set.

Interactive

Supervised Learning Algorithms

0/1 done
36

Overfitting & Underfitting

The bias-variance tradeoff — the single most important concept in applied ML.

Interactive

Advanced Learning Algorithms

0/1 done
36

Diagnosing Bias and Variance

The two fundamental failure modes of ML models — high bias underfits, high variance overfits.

Interactive

Supervised Learning Algorithms

0/1 done
37

Regularisation — Concept

Adding a penalty for large weights — the elegant way to prevent overfitting.

Interactive

Advanced Learning Algorithms

0/1 done
37

Regularization and Bias-Variance

How λ shifts the bias-variance tradeoff — and how cross-validation finds the sweet spot automatically.

Interactive

Supervised Learning Algorithms

0/1 done
38

Regularisation — Math for Linear Regression

L2 penalty added to MSE; weight decay in the gradient update.

Interactive

Advanced Learning Algorithms

0/1 done
38

Establishing a Baseline Level of Performance

Why raw error numbers are misleading without a baseline — human-level performance as the anchor for bias-variance judgment.

Interactive

Supervised Learning Algorithms

0/1 done
39

Regularised Logistic Regression

Applying L2 regularisation to logistic regression — the production standard.

Interactive

Advanced Learning Algorithms

0/24 done
39

Learning Curves

How training and cross-validation error change as data grows, and what that tells you about whether collecting more data is worth it.

Interactive
40

Deciding What to Try Next, Revisited

How bias and variance map directly to the next engineering move, so you stop guessing and start debugging systematically.

Interactive
41

Bias, Variance, and Neural Networks

Why deep learning changed the old bias-variance tradeoff story and gave engineers a new recipe for improving models.

Interactive
42

Iterative Loop of ML Development

The real workflow of ML engineering: choose architecture, train, diagnose, refine, and repeat until performance is good enough.

Interactive
43

Error Analysis

Manual review of model mistakes to discover which error classes matter most and where engineering effort will pay off.

Interactive
44

Adding Data

Targeted data collection, augmentation, and synthetic data generation as strategic tools for improving model quality.

Interactive
45

Transfer Learning

Use a model pre-trained on a large related dataset, then fine-tune it on your smaller task to get strong results with limited data.

Interactive
46

Full Cycle of a Machine Learning Project

Training a model is only one stage; real ML systems also require scoping, deployment, monitoring, retraining, and MLOps discipline.

Lab
47

Fairness, Bias, and Ethics

Why ML engineers must think about harm, subgroup performance, and mitigation plans before and after deployment.

Interactive
48

Error Metrics for Skewed Datasets

Why accuracy becomes misleading on rare-event problems, and how the confusion matrix gives a more truthful view of model usefulness.

Lab
49

Trading Off Precision and Recall

How threshold choices change which rare events you catch, which false alarms you accept, and why F1 is a useful but incomplete summary.

Interactive
50

Decision Tree Model

A decision tree predicts by asking a sequence of feature-based questions, routing an example down branches until it reaches a leaf decision.

Interactive
51

Decision Tree Learning Process

How a tree is built recursively: choose the best split, partition the data, repeat on each branch, and stop when further splitting is no longer worth it.

Lab
52

Measuring Purity: Entropy

Entropy is the impurity measure that tells a decision tree how mixed a node is, with 0 meaning pure and 1 meaning maximally mixed in the binary case.

Lab
53

Choosing a Split with Information Gain

Information gain measures how much a candidate split reduces weighted entropy, allowing the tree to choose the most purity-improving feature.

Interactive
54

Decision Tree: Putting It Together

The full tree-building algorithm combines repeated split selection, recursive branch construction, and stopping rules into one practical training loop.

Lab
55

One-Hot Encoding of Categorical Features

How to convert a feature with multiple discrete categories into several binary indicators so trees and other models can use it cleanly.

Lab
56

Continuous-Valued Features

How trees handle numeric features by testing candidate thresholds and selecting the split with the highest information gain.

Lab
57

Regression Trees

Generalizing decision trees from class prediction to numeric prediction by minimizing weighted variance and predicting leaf averages.

Lab
58

Using Multiple Decision Trees

Why single trees are sensitive to small data changes and how voting across many trees improves robustness.

Lab
59

Sampling with Replacement

Bootstrap sampling creates new training sets by repeatedly drawing from the original set with replacement.

Lab
60

Random Forest Algorithm

Bagging plus random feature subsets per split yields more diverse trees and stronger aggregate performance.

Lab
61

XGBoost

Boosted trees focus sequentially on hard examples and are often top-performing on structured/tabular tasks.

Interactive
62

When to Use Decision Trees

Choosing between tree ensembles and neural networks based on data modality, iteration speed, interpretability, and transfer learning needs.

Lab