The Entropy Mastermind Game as a Tool for Understanding Human and Artificial Intelligence

Beat the computer and become a Mastermind in this game-based session about computational complexity, information theory and thinking in humans and machines.

Human-like computing is a field of research with the aim to apply principles of human cognition, perception and action to the design of efficient computer algorithms. Computers learn from big data and can perform complex computation, while humans learn from few instances and often use computational shortcuts. Yet, the computational power of machines is limited and combinatoric problems may blow up to a point where the computer is not able to solve them anymore. Surprisingly, humans do comparatively well on some of these problems. Thus, equipping computer algorithms with human-like cognitive properties can increase the efficiency of algorithms and save computational resources.

In this event we will play a game, Entropy Mastermind, which is based on the classic Mastermind game. In this game, which confronts you with a computationally complex problem, you have to guess a secret code which was randomly generated from a probability distribution by making queries and getting feedback.

We will discuss strategies to solve the game efficiently and present basic concepts of probabilities and entropy, computational complexity and the differences between human- and machine-like problem solving. From there, we will derive key ideas for the development of efficient algorithms based on a combination of human-like and computer-like computational principles.

PUBLIC GAME SESSION ON ZOOM.

This is a digital event. When the event is taking place, you get access via a link on this site. Add the event to your schedule and receive a reminder. The number of participants is limited to 300. First come, first serve! 

Suitable for children.