Mirai at the Insurance Data Science 2018

Insurance Data Science was held on 16 July 2018 at Cass Business School in London.

As every year, the conference was a success with many interesting talks, panel discussions, and good organization. The slides of conference talks are available online. The talks had a varying scope bringing people from different fields together. The highlights of the conference included both the key note speeches. Gareth Peters from the Heriot-Watt University presented improved models to predict increasing longevity. Eric Novik from Generable presented the power of Stan, a highly performant tool for Bayesian decision making. One of the most interesting regular talks was given by Oliver Laslett from Cytora. He presented how fairness can be taken into account in pricing models.

Pauli represented Mirai with the talk titled: “PnC reinsurance modeling using Python and TensorFlow”. In the talk we showed how to apply TensorFlow in reinsurance models (such as Excess of Loss and Surplus share) requiring large-scale loss simulations. In addition, we showed what are the benefits and drawbacks of the approach. The talk was well received and we had many interesting discussions during the coffee breaks and the conference dinner organized in the evening in the fabulous Ironmongers’ Hall.

Next year the conference will be organized at the ETH Zurich, where we will naturally participate. See you next year again!