Deadline: 15 April 2013
Open to: individuals currently registered as students (graduate or undergraduate) in an academic institution
Prize: USD $5,000 and travel grants to attend SIGMOD 2013 in New York, USA
Student teams from degree granting institutions are invited to compete in the annual ACM SIGMOD (Association for Computing Machinery’s Special Interest Group on Management of Data) programming contest. This year, the task is to implement a streaming document filtering system. The winning team will be awarded a prize of USD $5,000. Submissions will be judged based on their overall performance on a supplied workload. Teams with the top-performing submissions will receive travel grants to attend SIGMOD 2013 in New York, USA .
The general idea is to filter a stream of documents using a dynamic set of exact and approximate continuous keyword match.
Specifically, the goal is to maximize the throughput with which documents are disseminated to active queries. Whenever a new document arrives, the system must quickly determine all queries satisfied by this document. Each query is represented as a set of keywords, and each document is represented as a sequence of space separated words. For a document to satisfy a query it should contain all the words in the query. Queries will fit into the main memory of the machine. Three types of keyword matching must be supported: exact matches, approximate matches under an edit distance constraint, and approximate matches under a Hamming distance constraint.
There are many practical examples where such a system can be used. For example, if one thinks of tweets as documents, and queries as some hash-tags a user wishes to follow, then it is desirable to know which users should be notified when a new tweet arrives. Another example is subscriptions to news services, where a user wishes to be notified each time an article of interest is published. The problem becomes challenging when the query set is big and changes frequently.
Since users interested in the same area usually submit similar queries, one particular focus will be on the practical case where a large number of queries have similar words. The organizers will provide a code submission system that will allow teams to upload their code and run it on our test machine, along with a leaderboard that will show the relative ranking of each of the teams. You may read a DETAILED DESCRIPTION of the task HERE.
Teams must consist of individuals currently registered as students (graduate or undergraduate) in an academic institution. Teams can register on the contest site after March 1, 2013. Multiple teams from an academic institution may participate.
- Teams must REGISTER HERE before being allowed to submit (as a test) a 64-bits Linux library implementing the contest interface functions in C/C++. The test submission will be evaluated and its results will appear on the leaderboard HERE. The team institution and country will not be shown in the leaderboard if this is specified during team registration. Teams are allowed to make several test submissions.
- Qualifying teams must submit the source code of their final implementation before 16 April. This final submission will be evaluated after the deadline and the finalists will be chosen based on it. Teams can submit several times, but only the last final submission will be considered.
- A submission must contain only code written by the team or open-source licensed software. Source code from books or public articles is permitted, such that a clear note about the source exists.
- By participating in this contest, each team agrees to publish its source code in case it is selected as a finalist. A team is eligible for the prize if at least one of its members will come to present his work at the SIGMOD 2013 conference.
Apr 15, 2013 Submission deadline.
May 1, 2013 Finalists will be notified.
To stay updated or to ask technical questions please join sigmod2013contest Google group. For non-technical questions, please contact:
The contest is organized by the InfoCloud Research Group in King Abdullah University of Science and Technology (KAUST) in collaboration with the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL).
For more information, please see the official website HERE.