Agenda:

6:30 7: Networking & pizza

7 7:05: Intro to Capital Ones Center for Machine Learning

7:05 7:30: Explainable AI (description below)

7:30 7:55: Fuzzy Search Case Study (description below)

7:55 8:30: Networking

Explainable AI: Key Techniques, Academic Literature & Societal Implications by Melissa Louie, Principal Associate, Data Science, Capital One

This presentation will provide an overview of Capital Ones research into Explainable AI (XAI), an area of artificial intelligence focused on being able to produce more explainable deep learning models while maintaining a high degree of efficacy, with the broader goal of ensuring humans ability to understand, trust, & provide context around machine learning outputs. The presentation will begin with an overview of the two most commonly-used techniques in XAI, including LIME & Integrated Gradients, then touch on additional methods from the latest academic literature, & will end with an overview of the broader implications of XAI for society & for enterprises.

Case Study in Applying Fuzzy Search Models to Transaction Data by Keira Zhou & Sheenam Mittal, Data Engineers, Capital One

When you look at your credit card transaction statement, have you ever wondered what SBUX means, what store is it & where is it? In this talk, youll learn about what Fuzzy Search is, & hear an use case of how Capital One applies. The presentation will focus on how we pair Fuzzy Search with other Machine Learning algorithms to cleanse transaction statements & map back to specific merchants, such as extrapolating & mapping Starbucks from SBUX.