Data Science Challenge: The dust has settled


When the i-Know conference came to its end on October 12th, excitement rang through the hall as the final winners of the i-Know Data Science Challenge were announced. The first prize went to Martin Lampel. He was handed a hugely oversized cheque of EUR 1.000,– by Mr. Dirk Theuerkauf, who represented our sponsor T-Mobile Austria.

Christian Reisinger was able to claim the second place with his solution, winning a cash prize of EUR 500,–. The third place went to Peter Schiffmann and Alexander Pfleger, who went on to receive a prize of EUR 250,–. They were congratulated by Prof. Stefanie Lindstaedt, Head of Know-Center and Professor at TU Graz.


The i-Know data science challenge 2017, jointly organised by the Know-Center and the Materials Center Leoben, attracted 40 participating teams. The contestants applied a variety of different methods to predict segregation energy from grain boundary atom positions. The best results were achieved by either domain-agnostically applying neural networks or approaching the problem with domain-specific knowledge and models. Although most participants are based in Austria, we are proud to have teams from all over Europe and even as far as the USA and India in our roster.


The competition data set is now available for everyone interested below. Should you be able to beat the topscore of an MSE of 0.0033 for the validation data set, feel free to drop as an email with your approach and result at For exceptional results, we will provide you with a goodie! (given the distance for postal delivery is sensible).

Download dataset (0,7 KB)


We are looking forward to our next Data Science Challenge, which we hope will contain more of everything: more rewards, more participants and even more complex problems to solve!

Data Science Challenge: The top five


On October 1, 2017, 23:55 PM MEZ, the final submission period for the Data Science Challenge has ended.

The team behind the Data Science Challenge hopes, that the task at hand has been a challenging and entertaining one for our contestants. Overwhelmed by the response from the community, we never even dared to hope for participation of this magnitude!

Today, we announced the top five contestants, who have achieved the best results by applying their models to the validation dataset. These contestants are invited to the award ceremony during the i-Know 2017, to which they also won free admission.

The ranking among the top five teams, and therefore also the ones winning the prizes, will be announced live at the Data Science Challenge award ceremony on October 12, 2017 17:30 MEZ at the i-Know conference in Graz.


After the conference, the complete results as well as the full dataset will be distributed to all participants as well as on the i-Know website. We would also like to keep in contact with our newly built community and therefore try to set up a way to keep in touch.

Once again, the Data Science Challenge Team wants to thank everyone who helped us out and participated in this challenge! We hope to have you on board once again the next time!

Data Science Challenge: Intermediate Results


The final count is in: 40 contesting teams and individuals have committed themselves to participating in the first i-KNOW Data Science Challenge!  The registration period is now closed.


Up until September 20, 2017, participants were able to submit one set of data for the purpose of preliminary ranking. Without further ado, here are the current top 5 teams.
Team IDs are composed of characters taken from participants registered adresses.

Team ID Rank Mean Squared
Minimum Absolute Deviation Maximum Absolute Deviation Medium Absolut Deviation
OGMC 1 0,01767 0,00001 0,27886 0,09996
RI.P 2 0,02385 0,00580 0,41600 0,11271
NE.I 3 0,02817 0,00260 0,41263 0,12446
CIFN 4 0,02906 0,00519 0,49847 0,12618
MNSB 5 0,03128 0,00514 0,54324 0,12608


Until October 1, 2017, 11:59 MEZ (UTC+1) participants are allowed to submit their final solution.  The 5 top ranking entries will then be asked to provide their source code and/or demonstrate their solution during a live session with the jury and describe their approach, applied methods and tools. After verifying the results, we will officially announce the winners.


Data Science Challenge released, registration period extended


More than thirty contesting teams and individuals from all over the globe have so far heeded the call to participate in the first i-KNOW Data Science Challenge. In regards to the resounding response in the last couple of days, most likely causally tied to the end of the summer holidays, we have decided to extend the registration period to September  20, 2017. Head over to the announcement page to sign up in the nick of time!


On September 11, 2017 we sent out the detailed  task description as well as the data sets to the contestants that signed up until that point. Newly registered participants will receive the challenge package shortly after signing up.


Starting with September 11, 2017 contestants are able to work on developing their solutions. Up until September 20, 2017, 23:55 MEZ (UTC+1) we will accept one optional preliminary submission per participant (team or individual) in a predetermined format. The preliminary ranking will be released on the i-KNOW 2017 website.

Until September 30, 2017, 23:55 MEZ (UTC+1) participants are allowed to submit their final solution. The 5 top ranking entries will then be asked to provide their source code or demonstrate their solution during a live session with the jury and describe their approach, applied methods and tools. After verifying the results, we will officially announce the winners.


Open registration for the Data Science Challenge

From 10th July 2017 onwards, we will officially open registration for the Data Science Challenge. To register, please fill out and submit the form at Further details regarding the conditions of participation can be found at the end of this article.


The Know-Center is organizing the first Data Science Challenge of the year! Sponsored by T-Mobile Austria and co-organized by Materials Center Leoben we invite students and experts to compete as teams or individuals in this exciting challenge in the field of data science leading up to the annual i-KNOW data-driven future conference. After the announcement of the challenge on 11th September 2017 all participants can use the three weeks up to 1st October 2017 to submit their developed code and solution proposal.

The highlight of our challenge will be the award ceremony at this year’s i-KNOW 2017 conference in Graz.