3.0

Average Course Rating

This course develops a theoretical understanding of the large-scale anatomical and functional organization of the human brain. We will discuss, present, and write about primary literature in the areas of theoretical and computational neuroscience, with connections to machine learning and artificial intelligence. Principles to be explored will include: hierarchy; normalization; pattern completion; prediction; gradient-based learning; and compositional representation. We will consider the motivation for each of these computational principles; we will ask how successfully they organize the empirical data about our brains; and we will explore whether they are also observed in machine intelligence. Specific questions include: What are the functional benefits of a hierarchical organization in the human cerebral cortex? Does the neocortex express repeated functional motifs? How and why is pattern completion implemented in the human brain? Which kinds of learning can occur without supervision or reinforcement signals? In what ways are human learning and machine learning fundamentally distinct? Cal 1; Programming is not required, but students should be willing to engage with computational concepts. Course Prerequisites: a) AS.110.106 / Calculus I OR AS.110.108 Calculus I b) AS.050.101 / Cognition OR AS.200.211 / Sensation & Perception OR AS.080.105 / Introduction to Neuroscience OR AS.050.203 OR instructor permission.