Prerequisites: It will be useful to have familiarity with modern machine learning concepts (e.g. CS229) and basic neural network training tools (eg. CS230 and/or CS231n). Prior knowledge of basic cognitive science or neuroscience not required, we will do a gentle introduction to neuro concepts as they arise.
Structure: Class sessions will be a combination of lectures focusing on neuroscience and AI/ML background and exposition of recent works showing how the two subjects can be combined fruitfully, and some occasional coding demonstrations sprinkled in. There will be a series of guest lectures (both from Stanford and outside faculty) on topics of interest. Students are expected to attend the course every session and to participate by asking questions actively throughout the quarter. (Question participation is a key aspect of the course.)
Coding Assignments: There will be two medium-sized coding assignments due throughout the quarter. These are mostly just meant to encourage students get there feet wet with training brain-like neural networks and comparing them to neural data.
Final Projects: In the last half of the quarter, students will work on a final project related to the material in the class. The project can tackle a slightly new modeling approach to an existing neural dataset, or can reproduce an existing piece of work from the literature. We will guide students to several resources for open neural datasets and brain evaluations that they can use asthe basis for their project, or students can use their own data if they have it.
Grading basis: participation (25%), coding assignments (40%), final project (35%).
Office hours: Dan Yamins (5:15-6:15pm Wednesdays, in the Wu Tsai Neurosciences 2nd Floor lounge); Klemen Kotar (5-6pm Wednesdays)
Structure: Class sessions will be a combination of lectures focusing on neuroscience and AI/ML background and exposition of recent works showing how the two subjects can be combined fruitfully, and some occasional coding demonstrations sprinkled in. There will be a series of guest lectures (both from Stanford and outside faculty) on topics of interest. Students are expected to attend the course every session and to participate by asking questions actively throughout the quarter. (Question participation is a key aspect of the course.)
Coding Assignments: There will be two medium-sized coding assignments due throughout the quarter. These are mostly just meant to encourage students get there feet wet with training brain-like neural networks and comparing them to neural data.
Final Projects: In the last half of the quarter, students will work on a final project related to the material in the class. The project can tackle a slightly new modeling approach to an existing neural dataset, or can reproduce an existing piece of work from the literature. We will guide students to several resources for open neural datasets and brain evaluations that they can use asthe basis for their project, or students can use their own data if they have it.
Grading basis: participation (25%), coding assignments (40%), final project (35%).
Office hours: Dan Yamins (5:15-6:15pm Wednesdays, in the Wu Tsai Neurosciences 2nd Floor lounge); Klemen Kotar (5-6pm Wednesdays)