Call for WSDM Cup Proposals


Important Dates


Proposals Submission Deadlines August 10, 2022
Notification of the Decision August 30, 2022
Competition Starts TBD
Competition Ends TBD
WSDM Cup Workshops On-site at the Conference (Finalists must present their solution at the workshop to qualify for the prizes) February 27, 2023

Contact email: wsdmcup2023@gmail.com


Description


We are inviting industrial and academic institutions to submit proposals for hosting WSDM Cup 2023, a machine-learning competition-style event co-located with the WSDM 2023 main conference at Singapore. Since 2016, WSDM cup has become one of the premier Data Mining competitions that attracts world-class experts to solve real-world and industrial-scale problems.

This year, we are looking for strong proposals that demonstrate novel and interesting challenges, with a broad outreach of the WSDM community, in particular those that differ from previous challenges.

Previous challenges have included a number of diversified tasks that aim at solving real-world problems, such as music recommendation, knowledge-base completion, fake news predictions and so on, from a variety of partners like Google, Amazon, iQiYi, Microsoft, Wikimedia, Adobe, ByteDance, Spotify, Baidu, and more. You can find task specifications, and associated results below for inspiration: 2016, 2017, 2018, 2019, 2020, 2022.


Requirements


We welcome submissions that meet the following criteria:

  • A clear description of the problem. Please make sure to clearly justify the novelty and importance of the problem, and why it is suitable for the WSDM community as well as the broader Data Mining community.
  • A problem should be challenging yet manageable. There should exist some room to improve from a baseline method provided by the organizer, and the task can be done within three months.
  • A clear description of the data. The data should be secured by the first day of competition and the test set should be confidential. Please describe the input and expected output of the problem. Including examples of data in the proposal is highly encouraged.
  • Accessibility. Non-domain experts (e.g., machine learning researchers) can participate in the competition without having prior knowledge of the problem. To this end, the organizer should clearly state the important prior knowledge in the proposal.
  • Metrics and evaluation methods. Please demonstrate how the submissions will be evaluated, with some clear definition of the metrics. A baseline method should be provided to assure non-trivial results can be achieved. We welcome you to introduce new metrics that are important in practice but haven’t been used in standard competitions before.
  • The platform used for the competition. You are welcome to use standard competition platforms like Kaggle, but feel free to use your own. Should you choose to use your own platform, please include a clear description of the platform in your proposal, and make sure it can be world-wide accessible.
  • Timeline and awards. The proposal should include the start and end time of the competition, with a detailed timeline for each competition stage. The number of final winners and the awards should also be clearly described.
  • The submitted proposal should be the version that you can publish on the competition's webpage after minor modification.

Please send your proposals in PDF format to wsdmcup2023@gmail.com by the submission deadlines described above.


WSDM Cup Chairs


Yang Song

Yang Song

Kuaishou Technology

Makoto Yamada

Makoto Yamada

Kyoto University, RIKEN