Learn Machine Learning
How a computer learns rules from examples, and the models that do it. Eighteen lessons start with features, labels and why the data is split. Then come the kinds of learning, the classic models from straight lines to forests, and neural networks and how they train. The last ones cover how to stop a model from memorising, and the libraries it is built with. It stops at building a model. Scoring one in depth and running one for real users are subjects of their own.
Not started
0 of 18 lessons done
Level 1 · How a machine learns
score0160/13 wordsWhat a model learns from, why the data is split, and the two ways a model can miss.
- Machine Learning BasicsWhat machine learning is, and how a model learns rules from examples. What features, labels and weights are, shown with house prices.11 min5 terms
- Training and Test SetsWhy data is cut into a training set, a validation set and a test set. How leaked answers make a model look better than it is.11 min4 terms
- Overfitting and UnderfittingWhy a model can be too simple, or learn its data too closely. How the gap between two scores shows it, and what bias and variance mean.12 min4 terms
Level 2 · Getting the data ready
score16280/6 wordsHow to turn raw data into features. Then how to get enough good examples, each with the right label.
- Feature EngineeringHow raw data becomes inputs a model can use. Make new features, one-hot encode the words, and put the numbers on one scale.12 min3 terms
- Data Labeling and AugmentationWhere labels come from, and what wrong ones cost. Why a rare class fools a model, and how changing the examples you have makes more of them.12 min3 terms
Level 3 · Kinds of learning
score28400/7 wordsWhere the right answers come from. Labels, no labels, a reward, an old model, or your own phone.
- Supervised and Unsupervised MLWhere a model gets its right answers. Supervised and unsupervised learning, and the two kinds that sit in between them.12 min4 terms
- RL and Transfer LearningThree more ways a model can learn. RL learns from rewards. Transfer learning starts from an old model. Federated learning trains on phones, and the data stays put.12 min3 terms
Level 4 · Classic models
score40660/20 wordsThe models that run most table data. Lines, neighbours, trees and groups, and how to tune them.
- Classification vs RegressionThe two main jobs of supervised learning. Classification picks a group and regression guesses a number. Then the simplest model for each.12 min4 terms
- k-NN, SVM and Naive BayesThree classic ways to sort examples into groups: by the nearest neighbours, by the odds, or by the widest gap. Plus how suggestions on film sites work.13 min4 terms
- Decision Trees and ForestsHow a decision tree asks yes or no questions, and why many trees beat one. Random forests, gradient boosting and XGBoost, the leaders on tables.13 min5 terms
- Clustering and PCAThree jobs with no answer key. Group like rows with k-means, squeeze many columns into a few with PCA, and spot the odd case out.13 min5 terms
- Hyperparameter TuningThe settings you pick before training, from tree depth to k, and how to search for the best ones on the validation set. Grid search and random search.11 min2 terms
Level 5 · Neural networks
score66880/21 wordsHow a network is built and how it learns. Then the designs for photos, for order, and for making new things.
- Neural Network BasicsHow a neural network is built from neurons in layers, joined by weights. What activation functions and softmax do, and what an MLP is.13 min7 terms
- Gradient Descent and BackpropHow a neural network learns. A loss score, backpropagation to find each weight's share of the error, and gradient descent to fix it, batch by batch.14 min7 terms
- CNNs, RNNs and LSTMsDeep learning and the network designs behind it. Why photos suit a CNN, why things that come in order need an RNN, and how an LSTM remembers longer.13 min4 terms
- GANs and Diffusion ModelsNetworks that make new things. How a GAN trains a maker against a judge, how diffusion turns noise into a picture, and what an autoencoder does.12 min3 terms
Level 6 · Better models, real tools
score881000/6 wordsHow to stop a model from learning its data by heart. Then the tools real models are built with.
- Regularization and DropoutWays to stop a model from learning its data by heart. A cost on big weights, neurons switched off at random, and stopping at the right epoch.12 min3 terms
- scikit-learn and PyTorchThe free Python tools most models are built with. scikit-learn for classic models on tables. PyTorch and TensorFlow for networks.12 min3 terms
Words in this course
The 73 glossary entries this course teaches, in the order it teaches them. Open any of them on its own.
- Machine Learning
- Training Data
- Feature (Machine Learning)
- Label (Machine Learning)
- Model Parameter
- Training Set
- Validation Set
- Test Set
- Data Leakage
- Overfitting
- Underfitting
- Generalization
- Bias-Variance Tradeoff
- Feature Engineering
- One-Hot Encoding
- Feature Scaling
- Data Labeling
- Class Imbalance
- Data Augmentation
- Supervised Learning
- Unsupervised Learning
- Semi-Supervised Learning
- Self-Supervised Learning
- Reinforcement Learning
- Transfer Learning
- Federated Learning
- Classification (Machine Learning)
- Regression (Machine Learning)
- Linear Regression
- Logistic Regression
- k-Nearest Neighbors (k-NN)
- Recommendation System
- Naive Bayes
- Support Vector Machine (SVM)
- Decision Tree
- Ensemble Learning
- Random Forest
- Gradient Boosting
- XGBoost
- Clustering
- k-Means Clustering
- Dimensionality Reduction
- Principal Component Analysis (PCA)
- Anomaly Detection
- Hyperparameter
- Hyperparameter Tuning
- Neural Network
- Artificial Neuron
- Layer (Neural Network)
- Weight (Neural Network)
- Activation Function
- Softmax Function
- Multilayer Perceptron (MLP)
- Loss Function
- Backpropagation
- Gradient Descent
- Learning Rate
- Optimizer (Machine Learning)
- Batch Size
- Epoch
- Deep Learning
- Convolutional Neural Network (CNN)
- Recurrent Neural Network (RNN)
- Long Short-Term Memory (LSTM)
- Generative Adversarial Network (GAN)
- Diffusion Model
- Autoencoder
- Regularization
- Dropout
- Early Stopping
- scikit-learn
- PyTorch
- TensorFlow