Courses

CS 182. Neural Basis of Thought and Language

Current Schedule (Spring 2008)

• CS C182: Jerome A. Feldman, T T 12:30-2:00P, 247 Cory [course homepage]

Description

This is an inherently interdisciplinary class with weekly problems ranging from linguistic analyses to programming of neural network algorithms. The final paper is an original synthesis in the style of the journal: Behavioral and Brain Sciences. Students are encouraged to work in teams and these tend to be mixed disciplines. Considerable attention is paid to the design of experiments - biological, psychological and computational. There is no specific lecture on professional responsibility or social responsibility, but both issues arise because of the importance of language understanding systems in society. The design aspect of the course lies in the detailed study of and experimenting with connectionist models and in two programming assignments that require design.

Prerequisites: CS 61B.

Course objectives: Learn about the computational structure of the human brain. Gain familiarity with neural computation and how it is modeled. Learn how language and thought can be understood as neural computations.

Topics covered:

• The Brain 1: Structure and function of the brain
• The Brain 2: Visual system and development
• Spreading activation
• Connectionist models
• PDP and structured connectionist models
• Backpropagation
• Computational models of learning
• Temporal binding
• Cognitive linguistics 1: prototypes
• Cognitive linguistics 2: image schemas
• Cognitive linguistics 3: image schemas, color, frame semantics
• Regier's model
• Bailey's model
• Events and causes
• Aspect 1: the inherent structure of events
• Aspect 2
• Metaphor 1: Narayanan's model
• Extending the NTL approach
• Temporal binding
• Grammar Acquisition