[Dbworld] SECOND Call for Participation in the FEIII Data Challenge
Ian Soboroff <ian.soboroff-R3+/[email protected]>
| Newsgroups | gmane.comp.db.dbworld |
|---|---|
| Message-ID | <[email protected]> |
Second Call for Shared Task Participation Financial Entity Identification and Information Integration Challenge https://ir.nist.gov/feiii/ ***** On January 16 (Martin Luther King Day in the USA), the SIGMOD Executive Committee decided to move the ACM SIGMOD/PODS 2017 conference out of North Carolina to a new, still undecided location. They will provide some details in the next few days about the new venue. ***** The challenge: identify relationships among companies. Given sentences from a 10-K regulatory filing that mention relationships with other companies, rank those in order of having important roles. Key dates: Data release: December 22, 2016 Task results due: March 17, 2017 Workshop: DATE AND VENUE COMING SOON! Challenge URL: https://ir.nist.gov/feiii/ Workshop URL: http://sigmod2017.org/workshops/ Introduction The Financial Entity Identification and Information Integration (FEIII) challenge series aims to provide interesting datasets to researchers at the intersection of finance and big data. These datasets have been partially curated and have sufficient data for exploration, but they may be noisy and incomplete. Each year we provide a specific SCORED evaluation task. We also provide some ideas for UNSCORED tasks and we encourage participants to define tasks that match their research interests. In Year One (2016), we posed an IDENTIFIER ALIGNMENT challenge: Given four databases of financial entities from four different sources, participants needed to find the entities in common across the databases. For Year Two (2017), we are continuing with the theme of IDENTIFYING and understanding RELATIONSHIPS among financial entities and the roles that they play in financial contracts. This dataset was created from 10-K and 10-Q filings. The task is a RANKED RETRIEVAL task, to identify relevant and interesting sentences in the filings that provide evidence for a specific relationship, described by a role keyword, between the filing financial entity and a mentioned financial entity. Shared Task Participants will receive a set of triples (mention of a financial entity; role keywords describing the possible relationship; context sentences around the mention and role keyword). The task is to order the triples, with a confidence score for each, indicating the system's confidence that the triple indicates the mentioned relationship between the filing company and the mentioned company. This is a tricky task because the roles themselves are not always well-defined, and the role keywords can have ordinary meanings outside of the regulatory jargon. Alternate Tasks We are also interested in participants that want to visualize role relationships and networks of companies, or other applications that support understanding a complex web of financial interactions among companies. Workshop The shared task and all other participation will report out at the DSMM 2017 workshop at SIGMOD 2017. The challenge website (https://ir.nist.gov/feiii/) has information on all due dates for both kinds of tasks. People Louiqa Raschid, University of Maryland Ian Soboroff, National Institute of Standards and Technology and collaborators at the Department of the Treasury Office of Financial Research and beyond. _______________________________________________ Please do not post msgs that are not relevant to the database community at large. Go to www.cs.wisc.edu/dbworld for guidelines and posting forms. To unsubscribe, go to https://lists.cs.wisc.edu/mailman/listinfo/dbworld