We are excited to open the call for proposals for hosting the WSDM Cup 2024, a premier machine-learning competition-style event that will be co-located with the WSDM 2024 main conference in Mérida, México. Since its inception in 2016, the WSDM Cup has attracted world-class experts to solve real-world and industrial-scale problems through innovative data mining solutions.
We are seeking proposals from both industrial and academic institutions that demonstrate novel and interesting challenges, with a broad outreach to the WSDM community. We encourage proposals that offer unique and distinct challenges that differ from those of previous years.
Previous challenges in the WSDM Cup have included a range of diversified tasks aimed at solving real-world problems, such as music recommendation, knowledge-base completion, fake news prediction, and more, from esteemed partners such as Google, Amazon, iQiYi, Microsoft, Wikimedia, Adobe, ByteDance, Spotify, Baidu, and others. For inspiration, we invite interested parties to review the task specifications and associated results from previous years: 2023; 2022; 2020; 2019; 2018; 2017; 2016.
We encourage all interested institutions to submit proposals and be part of this exciting event. Hosting the WSDM Cup 2023 provides an excellent opportunity to showcase your institution’s expertise, network with leading experts in the field, and contribute to the advancement of web search and data mining solutions for real-world problems.
We welcome submissions that meet the following criteria:
- A clear and well-justified description of the problem. We kindly request that proposals provide a clear and concise description of the problem with a strong rationale for its novelty and significance, as well as its relevance to the WSDM and broader data mining communities.
- A challenging yet manageable studied problem. The problem should be both challenging and feasible, allowing for improvements beyond the baseline method provided by the organizer, and it should be attainable within three months.
- A concise and well-defined explanation of the data. It is important that the experimental datasets are secured by the first day of the competition, and that they include a confidential test set. The proposal should clearly state the input and expected output of the problem, and providing data samples is highly recommended.
- Accessibility for non-domain experts. The competition should be designed to enable the participation of non-domain experts, such as machine learning researchers, who may not have prior knowledge of the problem. The organizer should be sure to provide a clear explanation of any important prior knowledge in the proposal to facilitate the participation of these individuals.
- A clear description of evaluation metrics and protocols. In order to ensure that the submissions are evaluated fairly, the proposal should define the evaluation metrics and methods clearly. The inclusion of a baseline method will help to establish a benchmark for non-trivial results. The proposers are encouraged to introduce new metrics that are relevant to the problem and have not been used in previous competitions.
- The platform used for the competition should be specified. The proposal should specify whether standard competition platforms like Kaggle will be used, or if the proposer will use a custom platform. If a custom platform is used, it should be described in detail, including its features and how it will be accessed globally.
- A comprehensive timeline for each stage of the competition. The proposal should provide a comprehensive timeline for the competition, including the start and end dates, as well as a detailed schedule for each stage of the competition. In addition to the timeline, the proposal should clearly state the number of final winners and the awards that will be presented.
- The submitted proposal should be the version that can be published on the competition’s webpage with minor modifications.
Please send your proposals to firstname.lastname@example.org in PDF format by the submission deadlines specified above.