Class Learning Objectives

CGS 4315 Intelligent Systems Design

Ability to select appropriate neural net design for given application problem
Ability to read and write formal rigorous mathematical statements
Ability to use Wolfe Conditions to Establish Convergence of Time-Varying Non-linear Optimization Algorithms
Ability to use multivariable calculus to characterize nonlinear objective function surfaces
Ability to use asymptotic statistical theory to make statistical inferences for non-standard neural net probability distributions on high-dimensional spaces
Ability to verify regularity conditions for applicability of asymptotic statistical theory
Ability to view neural nets formally as statistical pattern recognition algorithms
Ability to use Markov Random Fields for Analysis and Design
Ability to compute gradients and Hessians of objective functions
Ability to understand and apply basic notions of stochastic convergence
Ability to manipulate vector-valued discrete-time stochastic proceses
Ability to manipulate matrix algebra and calculus expressions
Ability to read and general formal statements in theorem format