Uni-Logo

Direktzugriff

    First place in Automatic Machine Learning Challenge

    27.05.2015 09:00
    Seven members of Dr. Frank Hutter's Automated Algorithm Design group and Prof. Martin Riedmiller's Machine Learning group won 1st place in the first phase of the Automatic Machine Learning (AutoML) Challenge.

    The challenge runs from February 2015 until January 2016 and is organized in 5 phases that offer a total of $USD 30.000 in prize money. Each phase contains five datasets from a broad range of machine learning applications, such as medical diagnosis, speech recognition, protein structure prediction, video classification, prediction of customer satisfaction, object recognition, or the prediction of toxicity and efficiency of new drugs. The study of such data sets by human machine learning experts is traditionally a lengthy and manual process. This competition now aims to automate this process and to make good predictions within minutes instead of months. Next to the main competition there is also a competition for human experts who can spend 2-3 months with the datasets.

    The IIF group focussed on the main challenge and won it in the first phase (prize money $1500). With their automatic tools they additionally won the 3rd place in the competition of human experts (prize month $500). More information can be found at: http://automl.chalearn.org/



    Benutzerspezifische Werkzeuge