Data Science Webinar: Interpretable Machine Learning 

13 September 2019 12 pm BST


 

Your details:

Our registration process uses cookies, by submitting this registration form you agree to our cookie policy.

(*) denotes required form field(s)

  Register

Data Science Webinar: Interpretable Machine Learning 

13 September 2019 12 pm BST


 

Overview

Interpretable Machine Learning
 
Sophisticated machine learning models have the potential for high predictive accuracy but their complexity may sometimes result in black box models, which, in some cases, may appear to be a trade-off between accuracy and interpretability.  For the actuary, the ability to articulate the outputs of a model is important and becomes crucial where models are used to inform important business decisions or where stakeholders need to understand the underlying dynamics of a system, and the impact of the results.
 
The presentation will cover
  • The importance of interpretability
  • What it means to have an interpretable machine learning model
  • Examples of approaches that have been used to provide interpretability  
  • A practical case study showing an example approach to explain model predictions.
  • Q&A
 
Background Material that could be useful for the session: 
Towards A Rigorous Science of Interpretable Machine Learning Finale Doshi-Velez∗ and Been Kim  https://arxiv.org/pdf/1702.08608.pdf

Presenters

Presenter
Michael Jordan
Actuary
Dupro Advisory
View Biography
Presenter
Rodwel Mupambirei
Actuary

View Biography
Presenter
Adriaan Rowan
Independent Actuarial and Analytics Consultant
i3Actuaries
View Biography