Presenter: Vijay Nair, Wells Fargo and University of Michigan, Ann Arbor

Abstract: The talk will start with a description of “my life as a quant” over the last 5+ years. As part of this, I will provide an overview of data-oriented modeling in a large bank, use of traditional statistical methods, and recent applications of machine learning (ML) algorithms. In regulated industries like banking, the model results have to be explained to various stakeholders such as customers, senior management, and regulators. This has been a focus of our own R&D work recently, and I will try to provide a critical review of different techniques for interpretability: post-hoc techniques, surrogate models, and self-interpretable models.

Dr Nair will also be giving a talk about the recruitment programs at Wells Fargo starting at 4:00 pm the same day, followed by a Q&A session. Please consider joining if you are interested.

For the interest of the audience, Dr Nair kindly made available a copy of his paper "Supervised Machine Learning Techniques: An Overview with Applications to Banking" that is appearing in International Statistical Review.