Deep Learning in Insurance – Theory and Practice

Event Type
Web Session
European Actuarial Academy

Two date event:

Start time: 8.00 am on Monday, 1st February 2021

End time: 3.00 pm on Tuesday, 2nd February 2021


Announcement from the European Actuarial Academy organiser: In computational science, deep learning probably is one of the most heralded techniques of present time and recent history, mainly due to its versatility and impressive achievements likewise. Indeed, applications of deep learning range from beating the (human) world champion of the highly complicated Go game to the promise of deploying self-driving cars in the near future, on a large scale and all over the world.

Deep learning (DL) pertains to the field of artificial intelligence and is great at extracting and mastering the often highly non-linear patterns of a given process, whatever this process might be. The only main requirement is the availability of a large amount of data that describes the behaviour of the process under different conditions and a truckload of computational power. However, since the price of data storage and the effort of sampling data has dropped dramatically over the last years, and since Moore’s law on the increase of computational power does even nowadays not show any signs of a slowdown, fitting deep learning models that are able to produce extremely useful predictions are a reality and this already for some years now.

In other words, the time is high to also deploy this amazing technology in the insurance industry! However, the methodological framework that underlies this amazing technology is somewhat different from the statistical one that we’ve all grown accustomed to (mainly through our general love for GLM models), and the computational horsepower, needed to merely fit these models, is of an order of magnitude higher than the one needed to fit the classical statistical models.

The practical sessions will make use of Keras, Tensorflow and R(Studio). Guidelines on how to install these tools on your own laptop will be provided several weeks prior to the beginning of the web session. As no practical hands-on exercise lesson will be provided, installation of these tools on your laptop is not a must, yet when the applications of DL will be discussed, participants will have the choice to run the code in real-time her/himself on her/his own laptop during this part of the web session.

The main purpose of this online training is to get the participants acquainted with DL models. To this end, a healthy mix between theory and practice will be provided, however, it is important to note that some time will be spend to go through the theoretical foundations of neural networks and hence DL, as the inner workings of these models are a bit different from the ones of the classic statistical models. Note that quite a bit of time will also spend on some real insurance applications, as to clearly illustrate how to use this amazing technology in practice. More specifically, applications on technical pricing in non-life and text analysis will be provided.

The registration fee is € 640 plus 19% VAT until 1 December 2020. After this date, the fee will be € 790 plus 19% VAT. Click here to make a reservation 


Monday, 01 February 2021

Introduction & welcome EAA

Introduction to DL models


High-level view: dissecting a DL model


Deep dive: loss functions & backprop


Applications in insurance: technical pricing non-life


Tuesday, 02 February 2021

Deep dive: backprop & SGD


Extra topics: regularization and RNN & CNN


Applications in insurance: text analysis


Applications cont'd & Q&A

Concluding remarks, closing of seminar (EAA)

Robin Van Oirbeek
Biographical details

Robin Van Oirbeek
Robin Van Oirbeek, after having worked as a statistical/actuarial consultant for different companies, is now working as Lead Data Scientist at Allianz Benelux. He is also an invited lecturer at the University of Antwerp (UAntwerpen) and at the University of Louvain-la-Neuve (UCLouvain). He uses R, amongst others, on a daily basis and this for around 15 years now.