People’s intuitive domain theories generate hypothesis spaces for concepts that could explain the features of objects that they observe. These hypothesis spaces can then be used to dramatically speed up learning, enabling people to generalize new features from very few examples. (Image by Prof. Joshua Tenenbaum.)
This course considers computational models of some of the core structures of human cognition: concepts, causal relationships, word meanings and intuitive theories. Class meetings mix lectures and discussion, covering both the necessary cognitive science and computational background and confronting state-of-the-art research questions.
An introduction to computational theories of human cognition. Emphasizes questions of inductive learning and inference, and the representation of knowledge. Project required for graduate credit. This class is suitable for intermediate to advanced undergraduates or graduate students specializing in cognitive science, artificial intelligence, and related fields.
Prerequisites: A course in cognitive science, and a course in probability or statistics.