Student teams from degree-granting institutions are invited to compete in the annual SIGMOD Programming Contest. This year, the subject of the contest is Hybrid Vector Search in high dimensional space. Teams' submissions will be judged on their performance on a 10-million-scale dataset.

The winning team will be awarded a prize of $2,000 (USD), and there will be an additional prize of $1,000 (USD) for the runner-up. Prize money is donated by Microsoft.

This year's contest is brought to you by the DB Group at Tsinghua University and the DB Group at Rutgers University. The organizing team is made up of Guoliang Li (co-chair), Dong Deng (co-chair), Jiayi Wang, Jiachi Wang, Ruilin Hu and Mengtong Xu.

For this year's contest, the task is to build an index for a set of vectors with attributes for answering hybrid vector queries. The index is required to efficiently find approximate k nearest neighbors of a given query vector under one given similarity measure, such as Euclidean distance, while satisfying the constraints on non-vector attributes.

For this task, you are asked to construct an index for one dataset, which contains a set of high-dimensional vectors with a categorical attribute and a timestamp attribute. The distances between vectors are measured by Euclidean distance, the categorical attribute is discretized into integers and the timestamp attribute is normalized into floats between 0 and 1. The challenge here is to answer hybrid vector queries accurately in a limited time.

More details about this year's task can be found on the Task page.

18 March 2024	New site up. Contest requirements specification and datasets available.
20 March 2024	Team registration begins. Leaderboard available.
3 May 2024	Final submission deadline (Eastern Standard Time).
6 May 2024	Finalists notified.
9-15 June 2024	ACM SIGMOD/PODS 2024 Conference.

Ask questions and stay up to date by joining the ACM SIGMOD 2024 Programming Contest Google Group or contact us at sigmodcontest24@outlook.com.