6 out of 200: Predicting a persons mood
Meet Christina Katsimerou in this Q&A series with 6 out of 200 computer scientists and mathematicians participating at the 3rd Heidelberg Laureate Forum, August 23–28, 2015. 26 Laureates (Abel Prize, Fields Medal, Nevanlinna Prize, Turing Award) will attend the forum together with them. For a full week Heidelberg in Germany will be the hot spot of mathematics and computer science.
Name? Christina Katsimerou
Where are you based? My research passion brought me to Delft, a picturesque town in the Netherlands. The campus of Delft University of Technology is getting more and more international year by year, offering equal opportunities to foreign and native students for research, socializing, networking, innovation, brainstorming, entrepreneurship, etc. Nevertheless living abroad makes me miss my friends, family and the Greek lifestyle. I feel quite lucky, though, because I did not abandon Greece out of necessity, but because of an interesting research opportunity. For the future I am optimistic to return back to my beautiful warm home country.
What is your current position? PhD candidate.
What is the focus of your research? Predicting the mood of a person from video.
Why did you become a computer scientist? The reasons why I chose to study CS were different from the reasons why I ended up loving it. I was actually never geeky, nor anyone in my family. I didn’t even own a computer until I was 18 years old. To figure out what to study was very difficult, because I had multiple interests, like Psychology, History or Math. My high school teachers convinced me that a CS degree will grant a very good job (partly true), and being good in math I have nothing to be afraid of (wrong), and that I can always combine it with a different study (true). So, after the intimidating first semesters, when I had to learn at the same time how to open a text file and program in C++, the most exciting years followed. They unveiled the spectrum of domains where CS could be applied: programming games, interfaces for controlling the computer with brain waves, diagnostic tools to help physicians, just to mention a few. As a Computer Scientist, I know I have the flexibility to choose my playground among all these application fields and the power to shape their future.
Anything like a favorite project? My favorite project is definitely my PhD project in Artificial Intelligence, trying to fight depression in elderly patients by making computers able to recognize their mood. This is a quite broad and interdisciplinary project with a combination of CS and Psychology. In the project I have to deal with quite intricate problems, as mood is expressed very subtly, and this is why it is often even difficult for other people to guess correctly someone’s mood. To crack the problem, I needed to collect the right data that reflect how people perceive other people’s mood. Being aware that a big amount of data is key for a successful data-driven computational models, I collected my ground truth with crowd-sourcing, which allowed me to retrieve thousands of data-points in only a few days. Based on this data and on psychological literature, I trained with an Artificial Neural Network a model that reads the expressions of the person over time and after a given timespan it predicts the person’s mood.
Whenever I describe my research to people, I often get the question: “Computers see only black or white, but there is no correct answer for mood. It is so subjective!” Well, in reality, I am making algorithms inspired by human empathy, so they predict the mood as subjectively as people do.
All these findings will be implemented in a prototype product. The goal of this project is to create an adaptive system, able to recognize the mood of an elderly in a care center room, and regulate the lighting accordingly, to alleviate the negative mood. This gentle and concrete purpose gives me additional motivation to see my work applied.
Any other interests? Besides my main research path, I have a lot of interests more or less relevant to science. The relation of my PhD topic to assisting elderly, brought me closer to the broader social and demographic problem of aging society. Together with 24 students from five engineering universities of Europe, I participated in a brainstorming yearly module for innovative solutions, which could alleviate the burden of this fragile part of the population and ensure a more sustainable society in the future.
Further away from research, I love dancing, and in particular Argentinian tango. There is a thought of turning my interest into an app to help people learn dancing at home, but at the moment I am solely enjoying dancing as a physical activity.
What do you expect from HLF? I think HLF is a great initiative to bring together international students, researchers and rewarded scientists. Admiration is a driving force for inspiration. I am sure that in the end of this forum, I will carry plenty of memories of inspiring stories. And of course, I am looking forward to meeting a few “nerdy” friends from all around the world!
Wish you an inspiring time in Heidelberg!