Industrial Challenge

We invite researchers to participate in a machine learning prediction challenge, where the task is to device a prediction model for judging whether or not a vehicle faces imminent failure of a specific component or not. The dataset consists of data collected from heavy Scania trucks in everyday usage. The system in focus is the Air Pressure system (APS) which generates pressurized air that are utilized in various functions in a truck, such as braking and gear changes. The datasets’ positive class corresponds to component failures for a specific component of the APS system. The negative class corresponds to trucks with failures for components not related to the APS system. The data consists of a subset of all available data and has been selected by experts from SCANIA.

The top contestants’ papers will be published in the IDA 2016 SPRINGER proceedings, and the authors be offered an oral presentation at the conference.  SCANIA, who supplied the data, will sponsor two prizes of a total of 5000SEK (approx. 450EUR). These two awards will be given to:

  • The best performing model: to be awarded to the authors/creators of the prediction model that achieves the lowest cost (see the assessment method here) of all participants on the given test set;
  • The most innovative approach: to be awarded to the most innovative approach for addressing the challenge.

Note that these awards are not mutually exclusive so one approach could win both prizes.

Two papers will be selected for inclusion in the conference proceedings (based on the above criteria). Additionally, a few more of the received submissions will be invited to participate in a poster session.

Submission instructions:

  • Deadline: July 1 2016
  • The contestants should submit two files:
    • A file (in TXT format) containing the predictions on the provided test set (look at the detailed instructions below for more information).
    • A short paper (in PDF format) where the overall approach for creating their proposed predictive model is described. It is encouraged that this paper is written in a  clear and inspirational style containing a crisp motivation of the authors’ choices when developing the proposed solution. The submission should be at most 5-pages single-column, following the formatting instructions of the regular paper submission of IDA 2016. For more information regarding formatting, please refer to the paper submission page.
  • The submission material should be sent via e-mail to: ida2016challenge-result@dsv.su.se
  • IMPORTANT: Only one submission is allowed per first author
  • The datasets can be found here: Training and Testing datasets (updated)
  • Detailed instructions and dataset description (updated)

For any questions please contact: ida2016challenge@dsv.su.se