⚙️
← Owens AI Institute

Machine Learning & Deep Learning

Data, training, neural networks — trained hands-on in your browser.

Elementary

Machines That Learn

Discover what it means for a machine to learn — getting better by practicing with examples, not just following fixed instructions.

  1. 1What Does It Mean to Learn?
  2. 2Machines Can Practice Too
  3. 3Programmed vs. Learned
  4. 4Learning From Examples
  5. 5Getting Better Over Time
  6. 6The Same Machine, New Tricks
  7. 7When a Machine Learns Something Wrong
  8. 8You Practice, Machines Practice
  9. 9Spot the Learning Machine
  10. 10Module Check

Data: A Machine's Examples

Data is just examples. Learn how machines are taught with pictures, sounds, and words — and what makes examples good and fair.

  1. 1What Is Data?
  2. 2Data Is Just Examples
  3. 3Pictures as Data
  4. 4Sounds and Words as Data
  5. 5Collecting Examples
  6. 6Good Examples, Bad Examples
  7. 7How Many Examples?
  8. 8Fair Examples for Everyone
  9. 9Data Detective
  10. 10Module Check

Teaching by Sorting

Teach a machine to sort things into groups using labels — and learn to check and fix its sorting.

  1. 1Putting Things in Groups
  2. 2What Is a Label?
  3. 3Labeling Examples
  4. 4The Sorting Machine
  5. 5Two Groups or Many Groups
  6. 6When Sorting Is Tricky
  7. 7Checking the Machine''s Work
  8. 8Fixing Sorting Mistakes
  9. 9Sorting Challenge
  10. 10Module Check
  11. 11Lab: Ml Pattern Finder

Predicting What Comes Next

Use examples and patterns to predict what comes next — the everyday superpower behind machine learning.

  1. 1Guessing With Clues
  2. 2Patterns Help Us Predict
  3. 3Predicting From Examples
  4. 4Everyday Predictions
  5. 5Predictions Can Be Wrong
  6. 6More Examples, Better Guesses
  7. 7How Sure Is the Machine?
  8. 8Predictions in Games and Apps
  9. 9Prediction Games
  10. 10Module Check

A First Look Inside Neural Networks

Take a first friendly look inside a neural network — tiny decision-makers, layers, and connections.

  1. 1Inspired by Brains
  2. 2Tiny Decision-Makers
  3. 3Passing Messages Along
  4. 4Layers Working Together
  5. 5Strong and Weak Connections
  6. 6How a Network Decides
  7. 7How a Network Practices
  8. 8Networks That See, Hear, and Read
  9. 9Neural Networks Aren''t Magic
  10. 10Module Check

Capstone Project: Build a Sorting Machine

Design a machine that sorts objects by a rule — and watch it “learn.”

  1. 1Project: Build a Sorting Machine

Middle

How Machine Learning Really Works

Open up machine learning — models, the training loop, the kinds of learning, and why learning beats hand-written rules.

  1. 1From Rules to Learning
  2. 2The Three Kinds of Learning
  3. 3What a Model Is
  4. 4The Training Loop
  5. 5Inputs, Outputs, and Examples
  6. 6Why Learning Beats Hand-Written Rules
  7. 7Machine Learning Already Runs Your Day
  8. 8The Limits of Learning
  9. 9ML or Not?
  10. 10Module Check

Data: The Fuel of ML

Data is the fuel of machine learning — features, labels, cleaning, train/test splits, and where bias hides.

  1. 1Datasets and Rows
  2. 2Features: What the Model Looks At
  3. 3Labels and Ground Truth
  4. 4Cleaning Messy Data
  5. 5Splitting Data: Train and Test
  6. 6Balanced vs. Skewed Data
  7. 7Where Bias Hides in Data
  8. 8Building a Tiny Dataset
  9. 9Data Workshop
  10. 10Module Check

Training a Model, Step by Step

Train a model step by step — predictions, error, iteration, accuracy, overfitting, and honest testing.

  1. 1Making a Prediction
  2. 2Measuring Error
  3. 3Learning From Mistakes
  4. 4Iterations and Improvement
  5. 5Accuracy and Other Scores
  6. 6Overfitting: Memorizing vs. Learning
  7. 7Testing Honestly
  8. 8Tuning a Model
  9. 9Train-a-Model Simulation
  10. 10Module Check
  11. 11Lab: Train And Test

Inside Neural Networks

Inside neural networks — neurons, weights, layers, activation, the forward pass, and how a network learns.

  1. 1The Artificial Neuron
  2. 2Weights and What They Store
  3. 3Stacking Neurons Into Layers
  4. 4Activation: Adding the Spark
  5. 5The Forward Pass
  6. 6How a Network Learns
  7. 7Deep vs Shallow Networks
  8. 8What Networks Are Good and Bad At
  9. 9Build-a-Network
  10. 10Module Check

Deep Learning in the Real World

Deep learning in the real world — vision, language, generation, why it took off, and its costs and failures.

  1. 1What Makes Learning 'Deep'
  2. 2Networks That See
  3. 3Networks That Understand Language
  4. 4Networks That Create
  5. 5Why Deep Learning Took Off
  6. 6The Cost of Deep Learning
  7. 7When Deep Learning Fails
  8. 8Using Deep Learning Responsibly
  9. 9Deep Learning Investigation
  10. 10Module Check

Capstone Project: Train a Mini Classifier

Run the full train-and-test workflow on data you collect.

  1. 1Project: Train a Mini Classifier

High School

The Mathematics of Learning

Learning seen clearly — functions, parameters, loss, generalization, the bias-variance trade-off, and optimization.

  1. 1Learning as Finding a Function
  2. 2Parameters and the Hypothesis Space
  3. 3From Data to a Model
  4. 4The Objective: Minimizing Loss
  5. 5Generalization — The Real Goal
  6. 6The Bias-Variance Trade-off
  7. 7Optimization as Hill-Descending
  8. 8Why More Data Helps (and When It Doesn't)
  9. 9Reasoning About a Learning Problem
  10. 10Module Check
  11. 11Lab: Gradient Descent

Supervised Learning in Depth

Classification and regression in depth — decision boundaries and the major algorithm families, from linear models to ensembles.

  1. 1The Supervised Learning Setup
  2. 2Classification vs Regression
  3. 3Decision Boundaries
  4. 4Linear and Logistic Regression
  5. 5Decision Trees
  6. 6k-Nearest Neighbors
  7. 7Ensembles — The Wisdom of Many Models
  8. 8Choosing the Right Algorithm
  9. 9Algorithm Showdown
  10. 10Module Check

Beyond Labels — Unsupervised & Reinforcement Learning

Learning without labels — clustering and dimensionality reduction — and reinforcement learning's agents, rewards, and policies.

  1. 1Learning Without Labels
  2. 2Clustering — Finding Natural Groups
  3. 3Dimensionality Reduction
  4. 4Anomaly Detection
  5. 5What Reinforcement Learning Is
  6. 6Agents, Environments, and Rewards
  7. 7Policies and Exploration
  8. 8Where Reinforcement Learning Shines and Struggles
  9. 9Designing a Learning Setup
  10. 10Module Check

Deep Neural Networks, In Depth

The real machinery of deep networks — the math of neurons, the forward pass, loss, gradient descent, backpropagation, and regularization.

  1. 1Why Go Deep
  2. 2Neurons, Weights, and Biases — The Math
  3. 3Activation Functions
  4. 4The Forward Pass, Formally
  5. 5Loss Functions
  6. 6Gradient Descent and the Learning Rate
  7. 7Backpropagation
  8. 8Overfitting and Regularization
  9. 9Training a Deep Network
  10. 10Module Check

Modern Deep Learning

Modern architectures — convolutional networks, sequence models, attention and transformers — training at scale, evaluation, and the frontier's limits.

  1. 1Convolutional Networks for Vision
  2. 2Sequence Models
  3. 3Attention and Transformers
  4. 4Training at Scale
  5. 5Transfer Learning and Fine-Tuning
  6. 6Evaluating Models Honestly
  7. 7Deploying and Monitoring Models
  8. 8The Limits and Frontier of Deep Learning
  9. 9Architecture Analysis
  10. 10Module Check

Capstone Project: An End-to-End ML Project

Carry one machine-learning problem from question to write-up.

  1. 1Project: An End-to-End ML Project