Neural networks have been tremendously successful in a variety of domains. Of course, they have also been extensively studied in actuarial research and applied in insurance companies. For example, an application which is highly relevant for life insurance is the modeling and forecasting of mortality rates.

While simple feed-forward neural networks are already very effective for a number of tasks, there are several techniques (such as embedding layers or ensembles) and specialized architectures (such as recurrent or convolutional neural networks) which are necessary to exploit the full potential of these powerful machine learning models.

Interpretability and explainability are often key requirements in actuarial applications. Successful approaches in this direction have been proposed and are, for example, based on combining neural networks with the more traditional generalized linear models in the so-called combined actuarial neural network or LocalGLMnet architectures.

Organised by the EAA – European Actuarial Academy GmbH.

This web session is directed at anyone interested in

• modern approaches to mortality modeling and forecasting using different kinds of neural networks,
• a broad overview of neural networks, either to get familiar with the topic or as a refreshment and expansion of existing knowledge.

Practical examples (using Jupyter, Keras and R) will be presented and made available to the participants. Instructions on how to install the necessary software will be provided to the participants in advance.

Technical Requirements
Please check with your IT department if your firewall and computer settings support web session participation (the programme Zoom will be used for this online training). Please also make sure that you are joining the web session with a stable internet connection.

The learning objectives of this web session are:

• Understand the basic functioning of neural networks
• Get an overview of more specialized neural network architectures, along with an understanding of how they work and what they are useful for
• Observe how neural networks are practically implemented, using the example of mortality forecasting, and be able to transfer this to conceptually similar applications

For achieving the third objective, a basic understanding of the R programming language would be helpful.
No previous knowledge of neural networks is necessary.

The language of the web session will be English.

Dr Simon Schnürch
Simon is working on calculation bases and pricing methods at Allianz Lebensversicherungs-AG (Stuttgart, Germany). During his PhD, which he obtained at Fraunhofer ITWM (Kaiserslautern, Germany), he investigated applications of machine learning methods, in particular neural networks, to mortality modeling.