This special topics graduate class explores the impacts of the recent advances in AI, particularly their use in automated decision making systems, from two complementary perspectives: ethical and economic. From the ethics of AI angle, we focus on three primary topics: fairness, privacy, and explainability. We explore the most common definitions and relevant algorithmic solutions in each area, as well as recent research advances in each direction. We then look at these same ethical questions, and also at the general use of AI, from an economic theory lens. This includes microeconomic questions such as gaming of algorithmic systems by human users, and macroeconomic questions such as the impacts of AI use on labor markets.

This graduate class is mainly about the fundamentals of random processes: how to model and analyze a collection of random variables indexed (usually) by time. The broad set of topics covered in this course are: • Review of probability; foundations of probability theory • Pairs of random variables; random vectors; random processes • Estimation; minimum mean square error estimation • Convergence of sequences of random variables; limit theorems • Discrete-time random processes; Markov chains • Continuous-time random processes; stationary random processes • Random processes through linear systems • Martingale theory.

This undergraduate course is mainly about learning the fundamentals of random processes: how to model and analyze a collection of random variables indexed (usually) by time. While we learn about random processes with a formal/theoretical approach, example application areas are mentioned for motivation throughout. The broad set of topics covered in this course are: • Review of probability: random variables, expected values, conditional/joint distributions/expectations • Paris of random variables; random vectors • Error estimation; detection • Random processes: definitions and basic properties • Markov Chains; Gaussian processes; Stationary processes.

This course explores the fundamentals of Markov decision processes and reinforcement learning algorithms. Topics covered includes the framework of Markov decision processes, exact dynamic programming, value and/or policy space approximations, and RL algorithms including Q-learning, policy gradient, actor-critic, and temporal differences methods.

This course exposes students to the fundamentals of nonlinear optimization theory and related algorithms. The broad areas covered are: • Optimality conditions for unconstrained and constrained optimization • Gradient and Newton methods, and convergence properties • Lagrange multipliers and associated necessary and sufficient conditions • Duality theory and convex programming • Algorithmic methods for unconstrained and constrained optimization, including steepest descent, Newton methods and variants, conjugate direction methods, interior point methods, and penalty and barrier methods

This course is an introduction to computer modeling and analysis of real-world systems under uncertainty. The goal of this course is to apply probability and computer simulation to model and analyze systems with time varying randomness. Such stochastic systems are commonly encountered in engineering, computer science, biology, finance, and public policy. Through this course, students gain experience in: • Understanding the role of stochastic modeling and simulation in system (re-)design and optimization • Planing and conducting data collection and analysis for discrete event simulation modeling • Formulating an appropriate simulation model for a system • Implementing the model as a computer program (in ARENA, as well as Excel, time-permitting) • Evaluating and interpreting the output of the simulation • Performing decision analytics by, e.g., making recommendations for system design and management based on the simulated model