Competition: Computer science teams determine the perfect first name for the child

Whether boy or girl - Kassel-based computer scientist Folke Mitzlaff experienced just how difficult it can be to find a suitable first name for his own child when the birth of his youngest daughter was due a year ago. "Unlike in the past, parents today include a variety of factors in the decision," says Mitzlaff, who is a research associate at the Department of Knowledge Processing at the University of Kassel. "Do we follow the trend or do we consciously disregard it?" he asks. At the same time, role models, role models or even cautionary examples are much more present through global networking than they were two generations ago." Some parents are also unsettled by reports of social and economic discrimination against children because of their first name, he said.

Mitzlaff, a computer scientist, applied knowledge processing methods to the problem. This resulted in the creation of the Nameling first name search engine(http://nameling.net) in early 2012, which provides recommendations to expectant parents. The search engine is now expanding its functions and is using a method that is also used by successful Internet department stores: "Customers who viewed this item also bought..." is often the phrase used there. Now it should be: "Parents who like Gustav also like ....".

Up until now, nameling.net was already able to do more than a classic first name encyclopedia: the search engine did not list first names alphabetically, but asked the user for a name they liked and suggested alternatives that appeared on the Internet in similar contexts. Now the Kassel-based computer scientists are going a big step further. "We can now personalize the suggestions," Mitzlaff explains. Users now receive recommendations according to search profiles - their own or those of previous Nameling users. So: many users who searched for Gustav may also have searched for Theodor or Emil or Adalbert and rated them positively. "After one year of operating nameling.net, we now have enough data for this," Mitzlaff says: "So far, more than 65,000 users from over 40 countries have used Nameling, in most cases expectant parents. In addition, Nameling now offers the option of displaying name suggestions for an entered list of several names. For example, users can search specifically for matching first names to the names of parents or siblings.

Exactly how the search engine is programmed is the subject of a scientific competition in which 17 research teams or individual scientists from around the world are participating. In collaboration with Jun.-Prof. Dr. Robert Jaeschke from the L3S research center in Hannover, Prof. Dr. Andreas Hotho from the University of Würzburg and his colleagues Stephan Doerfel and Jürgen Müller from Kassel, Mitzlaff won the bid to host the 15th "Discovery Challenge"; this is an annual international competition held as part of the "European Conference on Machine Learning and Principles and Practices of Knowledge Discovery," which will next be held in Prague in September 2013. As coordinator of the project, he is now calling on scientists from around the world to help expectant parents find a suitable first name.

To do this, the participating scientists are programming competing modules, each of which can take on the task of answering queries. Queries are then anonymously forwarded to the participants in Brazil, the USA or Germany and must be answered by the respective module in less than one second. The winner is the research team whose recommendation system provided the best suggestions, i.e. whose results were most liked by Nameling's users. For users, the competition is invisible behind the search engine. The contest has been running since August 1 and will continue until September 23.

Systems like Nameling cannot help expectant parents decide which is the right name for their child. However, they can help in the search and also offer alternatives to overly popular names. However, the scientists at the Knowledge Processing department are already thinking about new challenges in the search for first names: They want to investigate how opinion mining can be used to derive the subjective perception of names from public data from social platforms such as Twitter. In the future, nameling could then provide an indication, for example, if a name is associated with negative prejudices in a cultural or geographical context.

For more information, visit http://nameling.net

Photo: www.uni-kassel.de/uni/fileadmin/datas/uni/presse/anhaenge/2013/nameling_056.jpg
Caption: Computer scientists Jürgen Müller (back) and Folke Mitzlaff from the Department of Knowledge Processing at the University of Kassel. Photo: University of Kassel.

Contact:

Folke Mitzlaff
University of Kassel
FB 16 - Elektrotechnik/Informatik
Fachgebiet Wissensverarbeitung
Tel.: +49 561 804-6254

E-mail: mitzlaff[at]cs.uni-kassel[dot]de