Professional Documents
Culture Documents
{neuroscience}
Even if I love pasta, I will not order it food cup). Then, when the person experi- relevant to your current task were trying
at a Chinese restaurant, says Professor ences what actually happens in the world, to understand how we learn to attend only
Niv. The simple task of choosing an entre they can compute a prediction error - the to the aspects of the environment that are
on a restaurants menu requires your brain difference between the prediction and the relevant, in order to efficiently learn a task.
to take a myriad of considerations into ac- result. This prediction error enables people If a person focused on what table they were
count. Did I like this dish last time? Would to revise their predictions (and actions) in sitting at while deciding which entre to
this food be good at this type of restaurant? the future according to this error. order, that would not be efficient, as they
What would be the freshest option? Yet at Professor Nivs lab focuses on this con- would have to learn anew, at each table in
the same time, your brain must filter out all stant process of sorting the relevant from the restaurant, what dishes they like best.
other stimuli that dont affect that given de- the spurious a distinction that is central to In Professor Nivs words, her lab is study-
cision, like the color of the tablecloth. How real world learning. Professor Niv notes, In ing the central question: How do we learn
does your brain determine whats impor- the real world, the stimulus is not a simple what to learn about?
tant? light or a specific action. It is everything In order to study real-world learning,
In the lab, learning is usually studied in around you. How do learn efficiently about Professor Nivs lab tests how people solve
very specified situations with one stimulus, so many things at once? Luckily, you dont simple decision-making tasks. These tasks
for example: a study in which a rat receives have to, as most of whats around you is not are equivalent to small puzzles, such as
food for pressing on a lever. This type of
learning from feedback is called rein-
forcement learning. The main idea is that You are constantly making predictions,
people are constantly making predictions seeing the outcome, and revising your
(for instance, about how good the ordered
entre will be, or, for rats, whether pressing
predictions for the future.
the lever will result in food falling into the
4 innovation
PRINCETON UNIVERSITY life and mind
innovation 5