Meet the Winners of our CyberSec Game Challenge

Free Delivery on orders above £75 for our US and Europe customers
1 2

FuzzNet Labs

£30.00

Description

FuzzNet Labs is a competitive board game where players act as data scientists racing to build the best AI for recognizing cute animals in photos. You’ll design neural networks, train them with data, and test your models, while balancing risks like overfitting and messy datasets. Since neural networks are the core technology that powers modern Generative AI, mastering them in the game mirrors how breakthroughs in real-world AI are achieved.

Outcome

Gain a practical understanding of how AI and neural networks are developed, learning to balance data quality, model design, and training choices to achieve reliable results.

Run your own FuzzNet Labs session

Want to run your own FuzzNet Labs session? This downloadable pack created by Mike Novack is here to help. Ideal for workshops and training programmes, it gives you everything you need to run the game yourself and help participants grasp the concepts through hands-on learning.

Download the workshop toolkit

How to play FuzzNet Labs?

Watch this video guide to learn how to play and get started!

Prefer to read? Here’s how to set up and play

  • Goal

    Complete successful tests for all 5 animal models by designing, training, and testing your neural network. On your turn take one of the three actions: design, train or test your model.

  • Centre of table Set-up

    Place the Score Board in the middle of the table, face up on the side that matches the number of players.

    Place the dice, Node Tokens (x44), Data Tokens (x100), and Overfitting Tokens (x15) within reach of all players.

  • Player Set-up

    Each player receives a Player Board and Animal Tokens of a single matching color.

    Determine the first player and give them the First Player Marker.

    The first player places all of their Animal Tokens on the right-most column of their board in any order.


    All other players match this arrangement on their own boards.

    Each player places a Data Clean Token.

  • Design Model

    Add one Node Token to any empty node on your board.

    This creates or extends valid testing paths connecting animals for future tests.


  • Train Model

    Add two Data Tokens to nodes to improve your model’s chance of success.

    Each Data Token adds +1 to your dice total during a test.

    A node can hold a maximum of 3 Data Tokens - more causes Overfitting, blocking paths.

  • Test Model

    Roll three dice to test a specific animal using a valid path.

    The sum of your dice plus Data Token bonuses determines success.

    Passing all 5 animal tests completes your model.

Let's play FuzzNet Labs!

See the game in action. Watch this video to discover how to play.
Back to top