11-03, 13:30–14:10 (America/New_York), Radio City (Room 6604)
This talk explores how we create smarter, more reliable economic policies by using a technique called “Robust Control”. Robust control is a technique from control systems engineering focused on driving desired outcomes even when there is a lot of uncertainty about the circumstances and behaviors of the system these policies endeavor to regulate. These methodology assumes we've developed an incomplete structural model of a systems: parts of the system have well defined and enforceable rules, whereas others are models of phenomena outside of our control, such as user behavior. This talk reviews the application of a Robust Control informed workflow to select the parameters of a pricing algorithm. Code and data will be shared from the design and pre-launch tuning work and we will also use data to demonstrate how the real life deployed system did and did not match our models. The goal of this talk to demonstrate how practices from control engineering can be applied in data science applications, especially those where algorithms make decisions with intent to influence human behavior at the user population level.
The talk will start with a brief over view of concepts from control theory so that attendees who are not familiar with the field can follow along. After introducing the basic definitions, i will provide some context for when and why the methods classified as "robust control" are used, as opposed to other methods in control engineering (eg "adaptive control"). From there I will switch into example mode sharing sample code and data from a model-based systems engineering project that was completed in the past (using scientific python simulations to generate and evaluate scenarios). That project concluded in parameter recommendations which were incorporated into a production lending system, since that system was deployed its parameters have been tuned in order to keep the systems behavior within an acceptable range. I will conclude by talking about how the outcomes did and did not match the expectations of the designers, as well as the role the models we build played in the design and tuning of the parameters of this lending system.
No previous knowledge expected
Dr. Zargham is the founder and Chief Engineer at BlockScience, as well as a Research Director of a (nonprofit), The Metagovernance Project. Additionally, he serves on the Advisory Council at NumFocus. He holds a PhD in Electrical and Systems Engineering from the University of Pennsylvania with a focus on Optimal Dynamic Resource Allocation Policies.