Psych/Cog 411
Introduction to Neural Networks
Fall 2009
Instructor: David Vampola
Office: 113 Snygg Hall
Email: vampola@cs. oswego. edu
Telephone: (315) 312-2685
Office Hours: Tuesday 4:00 - 5:20, Monday and Friday 9:30 - 10:15
And by appointment
Main Course Objectives:
Students will:
gain an elementary understanding of the physical neuronal structure of the brain.
be able to understand and to run basic simulations of neuronal/brain functions. They will, furthermore, be able to understand the basic architectures and modeling assumptions of these simulations.
be able to link the processes that are created in the simulations to larger social
applications which involve “complex” situations.
understand what are the competing interpretations of connectionism and neural
networks, and they will be able to articulate the shortcomings of those approaches.
Prerequisites:
Students will be required to have/obtain an computer account through the university.
Course Texts:
The following required texts are available in the university bookstore:
1) O'Reilly, Randall and Munakata, Yuko. Computational Explorations in Cognitive Neuroscience,
(Cambridge, MA: MIT Press, 2000)
2) Lloyd, Dan. Radiant Cool: A Novel Theory of Consciousness (Cambridge, MA: MIT Press, 2004)
NOTE: We will also be using some simulation software in the course: details to follow!
Evaluation:
Quizzes: 10%
Mid-term exam: 15%
Final exam: 25%
Assignments: 15%
Term Project : 35%
PLEASE NOTE: GRADUATE STUDENTS WILL BE REQUIRED TO DO ADDITIONAL PROJECTS, AND THEY WILL MEET PERIODICALLY WITH ME FOR DISCUSSION.
Examinations and assignments will be based upon material covered in class, as well as readings from the textbook. It is expected that students will attend class regularly. No make-up exams will be given without a documented, legitimate excuse (such as a personal or family medical emergency). Assignments must be completed in a timely fashion. Late assignments will receive a reduction in grade.
Personal Responsibility:
It is expected that the student will assume responsibility for his/her performance in the course. Hence, it is incumbent on the student to bring any problems that he/she might be having in completing the required course work to the attention of the instructor as soon as possible.
Students with Disabilities:
Those students who need special consideration for whatever reason should notify the instructor at the beginning of the semester.
OUTLINE OF TOPICS
1. Modeling the emerging field of complexity
Fundamentals of Cognitive Modeling
2. Computational Cognitive Neuroscience - An overview
Reading: O’Reilly/Munakata, Chapter 1
3. Individual Neuron: Modeling Issues
Reading: O’Reilly/Munakata, Chapter 2
An excursion: The tlearn model of Plunkett and Elman
4. Networks of Neurons
Reading: O’Reilly/Munakata, Chapter 3
5. Hebbian Model Learning
Reading: O’Reilly/Munakata, Chapter 4
6. Error-Driven Task Learning
Reading: O’Reilly/Munakata, Chapter 5
7. Combined Model and Task Learning and Other Mechanisms
Reading: O’Reilly/Munakata, Chapter 6
8. Large-Scale Brain Area Functional Organization
Reading: O’Reilly/Munakata, Chapter 7
9. Perception and Attention
Reading: O’Reilly/Munakata, Chapter 8
10. Memory
Reading: O’Reilly/Munakata, Chapter 9
11. Language
Reading: O’Reilly/Munakata, Chapter 10
12. A Challenge and an Extension of Neural Networks: The "Case of Radiant Cool".
Reading: : Lloyd, Radiant Cool
13. Another model based upon neuroscience: "Neuron"
Applications of Connectionism and Neural Networks in the Social World
14. General Overview of Applications
Reading: Widrow et al “Neural Networks: Applications in Industry, Business and
Science”
15 Financial Applications of Connectionist Models
Reading: Wong and Selvi. “Neural Network Applications in Finance: A Review
and Analysis of Literature (1990 - 1996)”
16. Issues within Connectionist Approaches
Reading: O’Reilly/Munakata Chapter 12
Final Overview: The Network Paradigm and “Non-Linear” Thinking
NOTE: OTHER READINGS AND RESOURCES MAY BE ASSIGNED DURING THE SEMESTER.