"I wrote, deleted, rewrote"

Mr. Döttling, you have been awarded an ERC Starting Grant of almost 1.5 million euros for your research on "Laconic Cryptography". What does that mean?

Laconic cryptography is a term of art that I coined together with some co-authors a few years ago. In English, "laconic" is always used when someone needs very few words to express something. And that's what we're aiming for in cryptography, too. Here's an example: Imagine you want to program an artificial intelligence so that it can recognize a tumor on MRI scans. To do this, you need a lot of training data, i.e., real MRI images. We call the field "federated analysis." That means we use widely distributed data sources for machine learning. With this, you face a big challenge. That's because the images are held by medical institutions in different countries with a wide variety of data protection regulations. When you access it, they have to make sure that no personal data can be seen. The recipient should be able to use the data to learn how to recognize a tumor, but otherwise not be able to extract any other information. This can be done very well with today's cryptography, but you have to exchange data volumes over the network that are almost as large as the original data. For an MRI image, this is already in the gigabyte range, and you need thousands to train the AI. So the key questions are: How much computational time does it take to encrypt the information and what are the communication costs of doing so? In laconic cryptography, we want to minimize the data that is transmitted in such searches.

How did you actually get into cryptography?

What got me into cryptography was the beauty of the mathematics involved. I first came across it in 2007. I was a student then and had not yet completed my diploma. I was fascinated by the interplay between what I saw at the time as very abstract and advanced mathematics and an actual conceivable practical application. Most of the other advanced concepts of algebra that I had absorbed as a student were primarily about pushing the beauty of the theory further. But cryptography, in my view, combines two worlds. We study beautiful structures in mathematics, and we use them to create a social benefit. And that still excites me today. I then did my PhD at KIT and then continued to research cryptography as a postdoc in Aarhus in Denmark and in Berkeley in the US. Until 2018, I was a junior professor at Friedrich Alexander University Erlangen-Nuremberg and now I am here at CISPA.

So cryptography is the common thread running through your research career. How does the ERC Starting Grant help you now?

My research group and I have laid some foundations for laconic cryptography in the last few years. We were able to show that cryptographic methods with minimal data exchange are possible in principle. The protocols we have written for this work very well at the laboratory scale, but are not yet really practical for a real-world application. With the ERC Starting Grant, we will take up this challenge and find protocols that are, in principle, also conceivable for practical use. But for this, some fundamental things have to be completely rethought. That's why the project is scheduled to run for a full five years.

How much work actually goes into an ERC Starting Grant application?

First of all, you should keep one thing in mind: You don't just write an application for an ECR Grant. The application has to be interwoven with your own CV. All of your previous research is cumulative in it. You bring an ERC track record with you, as they say in English. In my case, it was "Laconic Cryptography", as it is a common thread running through my scientific career. So I spent several months thinking about the direction my project could take. Once the rough concept was in place, it was time to write, which took me three and a half months. I wrote, deleted, rewrote; you put something down on paper, give it to other people to read, and revise it. I think this process is enormously important. Because without feedback, you run the risk of getting bogged down in ideas that don't amount to much in the end. In all of this, of course, you always keep in mind that the chances of success are less than 10 percent. That's because only 400 of the approximately 4,500 applications were selected. In the first round, the jury reviews a five-page summary of the applications. After that, between 60 and 70 percent of the applicants are eliminated. If you make it through to the second round, you have to attend an interview. This normally takes place in Brussels. Due to the pandemic, this time it was a video conference. Each applicant must then appear before the panel and present his or her project in ten minutes. After that, they have to answer the jury's questions for another 10 minutes. You know that only 2 or 3 of the people on the panel have read through your application well. Now you have to convince the other 15 to 16 that you are worthy of funding. That involves a tremendous amount of stress. I hardly slept the night before and, like many others, memorized my presentation.

What does the grant mean for you personally and for your future scientific career?

Because the grant is so competitive, it also has a very high reputation. It is something of a seal of approval for researchers. For many people, it is a pound of gold that they can use to gain a permanent position. It was a little different for me. I already received my deferment a month before the winners were announced. Nevertheless, the grant is very important not only for my research, but also for myself. It shows that the scientific community recognizes and appreciates my research agenda.