These experiment demonstrations and simulations were designed and programmed by Rob Goldstone
Collective Behavior Experiments
A major activity of our lab is the study of collective behavior and the self-organized, dynamic patterns that groups of people make when they are trying to solve problems, make decisions, coordinate, search and use resources, and come up with new innovations. These demonstrations are most fun if you have multiple people playing at the same time.
- Crowd Estimation: Each player first tries to accurately estimate facts about countries on their own, then sees the estimates of others, and can decide whether to use these estimates to modify their own estimate. Relevant lab papers: 1 2 3 4 5 6 7 8 9 10.
- Group Binary Search: A server at Indiana University randomly generates a mystery number. Each player submits their own number. All of the players’ numbers are averaged together and compared to the mystery number. All of the players are given identical feedback on whether their group’s average was too low, too high, or just right, and each player decides for themselves whether and how to adjust their number for the next round. Your own guess shows up as a thick blue vertical line, everybody else’s guesses show up as red lines, and the group average is a dot (green: too low, red: too high, blue: just right). Relevant lab papers: 1 2 3 4 5 6 7
- Collective foraging: Players are rabbits foraging for morsels of grass that are placed in a 2D environment. Sometimes players see the other players and the grass, sometimes only the grass, and sometimes only the other players. How efficiently does the group allocate itself to resource patches over time? If there aren’t any human players to play with, the environment is populated with robot rabbits. Relevant lab papers: 1 2 3 4 5 6 7 8 9 10 11
- Wheel game: You will see 24 circles arranged in a bigger circle. You choose a circle by clicking on it. You earn a point for every person for whom you are exactly one position clockwise. You lose a point for every person for whom you are exactly one position counter-clockwise. What patterns do you see in how people’s choices are distributed? Relevant lab papers: 1 2 3
Create your own simple, two-frame movies, and then animate them to see how you naturally interpret the motion. Then, run a neural network computational model to see how it interprets the movie the same movie. Do you and the computational model interpret the motion in the same way?
A demonstration and computational model of binocular rivalry, continuous lag suppression, and 3-D perception via binocular fusion
To appreciate this demonstration, you’ll need to wear a pair of red/cyan glasses. Why not buy 50 for $8 and outfit your whole class, department, or friend network?
This online app lets you run yourself in a categorization experiment of your own design. You drag pictures from your computer into categories that you create. When you run the experiment, the computer will randomly present each of the pictures you supplied and log your response with feedback. After the last image presentation, the computer will display plots that show your accuracy and average correct response time for each category label. Check out the examples of experiments to see some of the possibilities.
Play the classic game of Nim, learn about the origins of combinatorial game theory, and prove why binary number representations lay hidden at the heart of the game (ball icons designed by Eleanor Goldstone).