Courses / DescriptionsEECS 435 - Neural Networks
Quarter OfferedFall : 2-2:50 MWF ; Lin
PrerequisitesEngineering Analysis 1 & 4, EECS 230 or EECS 231 (or equivalent), EECS 302.
CATALOG DESCRIPTION: Learning in one-layer and multi-layer feed-forward networks, recurrent networks and dynamical systems. Perceptrons, Hebbian learning, associative memories, Widrow-Hoff learning, backpropagation networks, radial basis function networks, competitive networks, counterprapagation networks, Grossberg network, Adaptive resonance theory, Hopfield networks, simulated annealing, Boltzmann machine
REQUIRED TEXT: TBD
COURSE GOALS: The goal of this course is to provide students with a basic understanding of the fundamentals and applications of artificial neural networks.
PREREQUISITES BY COURSES: Engineering Analysis 1 & 4, EECS 230 or EECS 231 (or equivalent), EECS 302.
PREREQUISISTES BY TOPIC: Linear Algebra, Differential Equations, Probability, Computer programming in MATLAB or C++.
DETAILED COURSE TOPICS:
Week 1: Introduction to artificial neural networks, neuron model and network architectures, perceptron learning rule
Week 2: Linear transformations for neural networks, supervised Hebbian learning, optimal linear associative memories
Week 3: Performance surfaces and optimum points, performance optimization, Widrow-Hoff learning
Week 4: Backpropagation learning algorithms, accelerated learning backpropagation algorithms
Week 5: Radial basis function networks, probability networks
Week 6: Associative learning
Week 7: Competitive networks, counterpropagation networks, Grossberg networks
Week 8: Adaptive resonance theory, stability
Week 9: Hopfield networks, bidirectional associative memories
Week 10: Simulated annealing, Boltzmann machine
COMPUTER PROJECTS: Projects on implementation of some neural network models and their applications to real-world problems will be assigned throughout the quarter.
Mid-term exam – 20%
Homework assignment – 50%
Final Exam – 30%
COURSE OBJECTIVES: When a student completes this course, s/he should be able to:
• Understand the mathematical foundations of neural network models.
• Design and implement neural network systems to solve real-world problems.