A question to follow, a project to make your own. Browse these possibilities, or bring an idea of your own.
When two things change together, how can you tell whether one causes the other? Build a small simulated world with hidden rules, then compare observing it with intervening in it. Design experiments to distinguish competing explanations within a limited budget. Investigate when more observations help and when you need a different experiment.
Link to this idea
How can a few measurements reveal an image or sound? Build a reconstruction laboratory with simple signals and simulated measurements. Vary the signal's structure, measurement count, and noise. Compare recovered signals with the originals, and find convincing reconstructions that are wrong. Explore which assumptions make recovery possible and when they fail.
Link to this idea
Simulate a toy game in which each independent fair coin toss adds 50% to a player's token balance on heads or removes 40% on tails. Compare the predicted average, the simulated mean, the median, and individual histories over many rounds. Explore how rare outcomes, changing stakes, or sharing resources affect the results.
Link to this idea
When does listening to others improve judgment, and when does it produce an echo chamber? Create simulated decision-makers who receive incomplete clues and exchange opinions. Vary the communication network, dependence on others, and whether they can remain undecided. Measure accuracy and agreement separately, and examine how your model's assumptions limit what you can conclude about real groups.
Link to this ideaAll 25 project ideas, the mentoring invitation, and next steps in one illustrated lookbook.
Mike Andrews P'28 · Independent ISP mentorship