AI Activities
Interactive simulations and hands-on labs for learning about Artificial Intelligence.
AI search and problem solving
Search algorithms in mazes
Explore BFS, DFS, Greedy best-first and A* and see how they spread through the maze, square by square. Ch. 2 · Fundamental search algorithms(pp. 51–65)Search algorithms in trees
Visualize fundamental search algorithms in a search tree and follow the search node by node. Ch. 2 · Fundamental search algorithms(pp. 51–65)Five in a row in a partially observable environment
Play five in a row without seeing all the squares to experience partial observability. Ch. 2 · Environments(pp. 43–45)Minimax with alpha‑beta pruning
Try the Minimax algorithm on a search tree. Follow the updates to the alpha-beta values during the search and see how alpha-beta pruning makes the search more efficient. Ch. 2 · Decision algorithms for games(pp. 66–74)Decision trees
Build a decision tree that tells venomous snakes from harmless ones. Test it on snakes it has never seen and compare it with the tree an algorithm builds. Ch. 2 · Decision trees(p. 47) Ch. 3 · Decision trees(pp. 91–94)Machine learning
Supervised learning
Neural networks
The neural networks lab
Build, train and explore a neural network. Follow the computations in each node, then let the network make predictions based on the input. Ch. 3 · Neural networks(pp. 112–117)Train a neural network
Follow a training round from beginning to end: forward pass, loss, backpropagation and weight updates.Advanced Ch. 3 · Neural networks(pp. 112–117)Visualizing decision boundaries
Place data points and compare how linear models and neural networks draw the boundaries between classes, and why neural networks are often needed. Ch. 3 · Neural networks(pp. 112–117)Computer vision with CNNs
Follow how an image-recognition model reads an image, from pixels through filters and feature maps to a classification. See how filters catch different patterns. CNN isn't covered in the textbookUnsupervised learning
Reinforcement learning
Q‑learning simulation
Train an agent to navigate a grid using reinforcement learning. Follow how the agent learns the best path by trying different actions and updating its Q-values. Ch. 3 · Reinforcement learning(pp. 124–130)NIM with reinforcement learning
Watch an AI system learn to play NIM from scratch. Train it over different numbers of games, follow the strategy forming in a Q-table and try playing against it. Ch. 3 · Reinforcement learning(pp. 124–130)Language models
AI and society
The bias lab
Train a recruitment AI on skewed historical data. Discover how discrimination arises and how hard it is to get rid of. Ch. 1 · Bias(pp. 31–32)The AI timeline: When was that?
Test your knowledge of AI history. Place events on a timeline and score points. Choose to practice solo or challenge a friend in two-player mode. Ch. 1 · The history of AI(pp. 14–19)Implement AI algorithms in code
Here you can try out how some AI algorithms work by running ready-made code blocks in Google Colab. You do not need to know how to program. It is enough to follow the instructions, make small changes and see how the result is affected. Note: these notebooks are in Swedish.
BFS and DFS in trees
Google Colab Swedish only Ch. 2 · Fundamental search algorithms(pp. 51–65) (opens in a new tab)Linear regression with gradient descent
Google Colab Swedish only Ch. 3 · Linear regression and gradient descent(pp. 96–101) (opens in a new tab)k‑nearest neighbours
Google Colab Swedish only Ch. 3 · k‑nearest neighbors (kNN)(pp. 103–105) (opens in a new tab)k‑means clustering
Google Colab Swedish only Ch. 3 · The k‑means algorithm(pp. 120–122) (opens in a new tab)
The textbook
AI – Teknik, samhälle och etik
All the simulations on this site were created as a digital companion to the textbook AI – Teknik, samhälle och etik (a Swedish AI textbook).
Explore the book at Liber