|
|
CS 564 : Fall 2001 Brain
Theory and Artificial Intelligence |
Tu
Th: 9:30-10:50 am; OHE 100
Instructor:
Prof. Michael A. Arbib;
HNB-03, (213) 740-9220, arbib@pollux.usc.edu. (Office hours: 11-12
Tuesdays, HNB 03.)
Teaching
Assistants: Erhan Oztop, erhan@java.usc.edu, Salvador Marmol, smarmol@rana
This course is for any graduate student who has been inspired by either of two great questions "How does the brain work?" or "How can we build intelligent machines?" The course is carefully designed so that students may come to the course with a background either in technology, psychology or neuroscience and learn the basics of the other disciplines needed to work on an interdisciplinary team modeling some specific brain mechanisms underlying primate behavior. Students will find the course both challenging and stimulating.
Prerequisites: Graduate standing; ability to program in C++ or Java, or strong background in neuroscience or the behavioral sciences.
Texts:
[TMB]
M.A. Arbib, 1989, The Metaphorical Brain 2: Neural Networks and
Beyond, Wiley-Interscience.
[NSLbook]
A. Weitzenfeld, M.A. Arbib and A. Alexander, 2000, NSL Neural
Simulation Language, MIT Press (in press). [http://www-hbp.usc.edu/_Documentation/NSL/Book/TOC.htm]
Supplementary
reading:
M.A.
Arbib, Ed., 1995, The Handbook of Brain Theory and Neural
Networks, MIT Press (paperback).
Michael
A. Arbib, and Jeffrey Grethe, Editors, 2001, Computing the
Brain: A Guide to Neuroinformatics, and the Project Team of
the University of Southern California Brain Project, San Diego:
Academic Press.
|
1.
8/28 |
Modeling
the Mirror System: Setting Goals for the Course [Proposal]
{Background: TMB Chapter 1} |
|
2.
8/30 |
Charting
the Brain 1 [TMB 2.4] |
|
3.
9/4 |
The
Brain as a Network of Neurons [TMB Section 2.3] |
|
4.
9/6 |
Visual
Preprocessing [TMB 3.3] |
|
5.
9/11 |
Adaptive
networks: Hebbian learning, Perceptrons; Landmark learning [TMB
3.4] [NSLbook] |
|
6.
9/13 |
Visual
plasticity; Self-organizing feature maps; [NSLJ] Kohonen
maps |
|
7.
9/18 |
Higher
level vision 1: object recognition {Background TMB 5.2} |
|
8.
9/20x |
Introduction
to NSL: modules; SCS schematic capture system; Maxselector
model[NSLbook] {Homework} (Marmol) |
|
9.
9/25 |
The
FARS model 1: Reaching, Grasping and Affordances [TMB 2.2,
5.3; FARS Paper] |
|
10.
9/27 |
The
MNS1 Model 1: Basic Schemas and the Core Mirror Neuron
Circuit [MNS paper] |
|
11.
10/2 |
Overview
of the Five Projects {Mirror Neuron Proposal] |
|
12.
10/4 |
NeuroBench
and the NeuroHomology Database (Oztop and Bota) |
|
13.
10/9 |
The
FARS model 2: [FARS paper] |
|
14.
10/11 |
FARS,
Synthetic PET, MNS and
Imitation [Synthetic Brain Imagingpaper] |
|
10/16x |
Midterm |
|
15.
10/18x |
Adaptive
networks: Gradient descent and backpropagation [TMB 8.2] (Oztop) |
|
16.
10/23x |
[NSLJ]
Backprop: a. How to run the model; b. How to write the model
[NSLbook] (Marmol) |
|
17.
10/25 |
Extending the FARS model to mirror neurons and language [First Project Report due] |
|
18.
10/30 |
Systems concepts; Feedback and the spinal cord [TMB 3.1, 3.2] |
|
19.
11/1 |
Control of eye movements [TMB 6.2] |
|
20.
11/6 |
Basal Ganglia and Control of eye movements - Dominey [NSLbook] |
|
21.
11/8x |
Reinforcement learning and motor control (Sethu Vijayakumar) |
|
22.
11/13 |
The MNS1 Model 2: Hand Recognition; Simulating the kinematics and biomechanics of reach and grasp |
|
23.
11/15 |
The MNS 1 Model 3: Modeling the Core Mirror Neuron Circuit |
|
24.
11/20 |
Dopamine and Sequence Learning[Dopamine and Planning paper] |
|
11/22 |
Thanksgiving
holiday |
|
25.
11/27 |
Abstract
models of Sequence Learning |
|
26.
11/29 |
Project
Reports 1, 2,3 |
|
27.
12/4 |
Project
Reports 4,5; Concluding Discussion |
|
28.
12/6 |
No Class |
One
mid-term and a final will cover the entire contents of the
readings as well as the lectures.
Students
will be organized into 5 groups, each working together on a
semester-long project.
The
final exam will cover all of the course, but emphasizing
material not covered in the mid-term.
Distribution of Grades:
NSL assignments and other homework: 25%; Mid-term: 20%;
Project 30%; Final Exam: 25%.
Spares: Material covered last year but omitted from the syllabus: Perceptual and motor schemas [TMB 2.2, 5.1 and 5.2]; A first neural network: Didday model of winner take all [TMB 4.3 (last part) and 4.4] ; The story so far: An integrative view; Stereoscopic vision [TMB 7.1]; Motion perception and optic flow [TMB 7.2]; Higher level vision 2: visual attention; Memory and Consciousness [TMB 8.3].