Illusions to Inference
PREREQUISITE: MATH-UA 121.
Human perception easily falls prey to illusions and biases. It is tempting to think of these as failures of our brain, but they might not be. In fact, many illusions are likely byproducts of smart solutions that the brain applies when having to interpret imperfect and incomplete sensory input. We will study a wide variety of well-known and lesser-known illusions (visual, auditory, tactile, vestibular, and multisensory) to develop the central concept of probabilistic inference, the notion that the brain constantly forms hypotheses about the world and tries to figure out which of them is most probable given the sensory input. We will make connections to other applications of probabilistic inference, including medical diagnosis,crime scene investigation, election forecasting, and data science.