Hotshot leaves Google for SDU

By Birgitte Dalgaard, bird@tek.sdu.dk, 1/1/0001

Almost 25 years ago, she packed her suitcases and moved to California because her research in computer science was too controversial for the Danish research climate. A job as professor and head of SDU Applied AI and Data Science at the Maersk Mc-Kinney Møller Institute has now made Elin Rønby Pedersen turn herself towards Denmark again.

- Instead of artificial intelligence, I very much prefer that we call it machine learning, says expert in human-centred computing and machine learning, Elin Rønby Pedersen:

- It's amazing what we can do with machine learning. We can stand on any street corner in the world and point the phone at a road sign, and the phone will tell us where we are and translate what is written on the signs. It's crazy, but is it intelligence? I do not know.

She got her no bullshit attitude in a world full of buzzwords, from world-famous anthropology professor Lucy Suchman, whom she first met in 1989 at Xerox PARC, Palo Alto Research Center in California. Back then, Elin Rønby Pedersen as a young and very technology-enthusiastic computer scientist first set foot in Silicon Valley.

- When I was having lunch with Lucy Suchman and her group, they asked me: What is it actually good for? Why do you think it saves someone's life? In this way, I was pressured to reflect on the fact that technology only provides value when it makes sense in people's daily lives, explains Elin Rønby Pedersen:

- It seems banal, but it is not in the world of technology.

The person behind data

For 14 years, Elin Rønby Pedersen has been employed as a researcher in machine learning within the health department at Google, and in the stronghold of the tech giants, she has succeeded in manifesting her basic rule that algorithms must be relevant and developed with respect to the people they are to help, and the reality that they will be part of.

- Recently, I was part of an interesting project that dealt with health workers in the Third World. The problem was that the data they collected about patients' illness and treatment was often inaccurate. The approach to solving the problem will usually be to give the health workers an app that controls that they do not make mistakes, says Elin Rønby Pedersen and continues:

- But since the 80's we have known that if you ask people to enter a lot of data that they have no interest in, then data will be flawed. That's what Jonathan Grudin from Microsoft said at the time.

- So what if we instead make data meaningful for health workers. Take their professionalism seriously and give them something in return. For example, in the form of the opportunity to provide better treatment based on the data collected. That is the point of view that I'm working on, and it is now a key theme in Google's research and development.

A Treasure chest of health data

Now Elin Rønby Pedersen has replaced Google with the University of Southern Denmark to take the position as professor and head of SDU Applied AI and Data Science at the Maersk Mc-Kinney Moller Institute. And she does not hide the fact that it is the very special historical collection of Danish health data, that has also drawn her to Denmark.

- It's a treasure chest of data. For a person with an interest in machine learning, it is like a candy store, says Elin Rønby Pedersen.

She is sure that with her skills in machine learning, she can turn the treasure chest of data into useful knowledge about the development and danger signals of many diseases.

- I don’t understand why Denmark is not the world champion in the development of health systems based on its completely unique gold mine of data, says Elin Rønby Pedersen.

She points out that she is really curious about what it takes to access data, or whether there is a lack of reasonable rules that ensures that the data collected doesn’t just lay dormant, but that it instead is used to improve patients' opportunities for receiving a quick diagnosis and treatment.

- I would like to see if we can come up with some reasonable criteria for when it is wise to use the data. Here lies an enormous untapped potential, which can create great value for Denmark, points out Elin Rønby Pedersen.