We invite industrial, government, and academic organizations to submit proposals to organize a Data Challenge for CIKM 2017. CIKM Data Challenges are analytics and programming contests based on datasets released by the organizations sponsoring the competitions. These contests will run for several months, with winners announced at the CIKM conference in November 2017.
Strong proposals will release interesting and novel datasets to the participants and propose tasks that are challenging, yet achievable within the expected timeline. To encourage broad participation in the challenge competition, tasks and data should be made publicly accessible where possible.
We will accept at most one Data Challenge in each of the six sectors of interest:
Specific Task Challenges involve a released dataset and a well-defined problem with clear evaluation metrics. Specific Tasks run on a data competition platform (e.g., Kaggle) that maintains a leaderboard for the participants. Semi-finalists are selected based on their standing on the leaderboard, and must then submit a 4-page report describing their algorithms and results.
Open Task Challenges involve a released dataset and two rounds of competition. In the first round, participating teams propose tasks by submitting a 500-word abstract. Teams selected in the first round are invited to submit an implementation or a demonstration of their respective proposed tasks in the second round.
Proposals should answer the following questions:
For a specific-task challenge:
For an open-task challenge:
The CIKM’17 AnalytiCup Co-Chairs are Xiaoli Li, A*STAR, Singapore; & Hady W. Lauw, Singapore Management University.
Specific-Task Challenges & Open-Task Challenges