Explainable Machine Learning

Explainable Machine Learning

Monday 27th April 2020 12 PM BST

Machine learning techniques have become more and more popular within the financial industry mainly because of

- the potential to capture complex interactions from data

- the potential for better predictive models than traditional statistical models

- the ability to capture non-linear interactions within a range of inputs

‍

Machine Learning techniques have been viewed as useful additions in the actuary’s modelling toolkit that could enable insurers to process and learn from more data

Nevertheless, these models - sometimes viewed as black box models - can sometimes be hard to interpret, audit and debug, subsequently making it harder to trust and use the outcome of the prediction resulting from these models.

Building on from Actuartech’s previous webinar on Interpretable Machine Learning we are pleased to bring another insights session on the topic of explainable Machine Learning - presented by Reacfin.

We will talk about some of the worries around delegating decisions to machines, and how to overcome some of these challenges. We will touch on some of the issues surrounding the trade-off between predictive power and explainability.

We will also expand beyond some of the technical concepts and explore ways to design and build these models in order to make them usable within stakeholder communication scenarios as well as make them suitable to meet professional conduct requirements and usable in the context of achieving fairness in insurance pricing.

During this webinar we will expand on some of the previous techniques highlighted and will introduce additional techniques that can be used in order to better understand and interpret machine learning models and results, showing why they need not be viewed as ‘black box models’.

We will build on this to help identify ways to obtain sufficient comfort in the models in order to make business decisions and to be able to explain the impact to stakeholders.

These interpretability tools make the use of ML techniques much more relevant in practice as it enables to benefit from their higher predictive power while understanding the drivers of the results; which is fundamental to adopt them and take relevant business decisions. As these new methods for investigating machine learning methods are developed, we expect insurers to grow more confident in using these methods and shift more towards models with higher predictive power.

The presentation will cover:

- The importance of interpretability

- What it means to have an interpretable machine learning model

- A non-exhaustive reminder of some useful machine learning techniques

- An introduction to machine learning interpretation tools

- How to make the most of machine learning techniques            

- Q&A

‍