Artificial intelligence on the rise

Whether Alexa, Cortana or Siri: Artificial intelligence is on the rise. Therefore ERGO CDO Mark Klein and some members of the Board of Management at ERGO started a conversation with the experts from the German Research Centre for Artificial Intelligence. Mark Klein reports in his blog on this dialogue.

Engaging with the German Research Centre

More and more companies are focusing on artificial intelligence. A few months ago Google, for example, said goodbye to ‘Mobile first’, which had been their focus before and went for a new line of approach with ‘Artificial Intelligence first’. Therefore, I was very pleased to discuss in Kaiserslautern artificial intelligence trends and areas to which it can be applied for ERGO with Professor Wolfgang Wahlster from the German Research Centre for Artificial Intelligence (DFKI) and Silke Lautenschläger and Thomasz Smaczny, my colleagues who are members of the Board of Management at ERGO.

From ‘robotics’ to ‘smart city’

Visit at German Research Centre for Artificial IntelligenceThe DFKI is one of the best research centres in the world: there are currently 480 highly qualified academics and 376 student assistants from more than 60 nations who are working on 180 research projects. About 70 spin-offs have their roots in the centre, which has created more than 3,200 new jobs. The centre works closely with industry regarding joint ventures. Companies include digital leaders such as Google, Microsoft and SAP, but also more traditional players such as BMW, Bosch and Airbus. The DFKI works in seven business segments, ranging from ‘robotics’ to ‘smart city’. ‘Smart insurance’ has not yet been included. This has shown me yet again that we, at ERGO, are pioneers in this field.

Fascinating: deep machine learning

I was particularly fascinated by the so-called ‘deep machine learning’ during the discussion, in which computers themselves acquire knowledge in artificial neural networks. This allows complex tasks to be automated and executed at very high speed by a computer – for example, the analysis of images: the computer recognises whether the image has a positive content or whether it has already been sent and has only been changed slightly. For example, fraud attempts can be averted.

Initially, it may come as a surprise that ‘machine learning’ is attracting so much attention at the moment: the technology is based on artificial neural networks, which, in theory, have existed since the early 1940s and were algorithmically tested in the 1980s. Since then, however, three things have fundamentally changed:

1. Today, much more digital data is available, which is necessary for training these larger neural networks. Sources for this are, for example, social media sites such as Twitter.

2. The development of Graphic Processing Units (GPU) makes it possible to carry out the numerous calculations in parallel and thereby to make neural networks much faster. Although GPUs were originally intended for electronic gaming, they are now in demand in all fields where artificial intelligence is gaining in importance.

3. The size of neural networks has increased dramatically in recent years. It depends, among others, on the number of layers in a network. Each layer consists of interacting nodes representing the finer information, while the layers in conjunction represent the broader insight. Previously, networks were built up to eight layers, today it is possible to have a network with more than 140. Thus, the networks become considerably more powerful.

With the interplay of these three factors, the developments in ‘deep machine learning’ are progressing rapidly. They offer many opportunities for ERGO to impress their customers – for example, in customer service, where we are able to process customer concerns faster. I am keen to take advantage of these opportunities together with my colleagues at ERGO. The future is now.

I look forward to receiving your comments, questions and suggestions.
Mark Klein