OKN: Open Knowledge Network:
Creating the Semantic Information Infrastructure for the Future[1]
Proposed Next Steps
Under the direction of the three co-leaders: Andrew Moore (Academic Sector), R.V. Guha (Private Sector), and Chaitan Baru (Government Sector), the following working groups have been proposed as a method of moving forward with the OKN project. If you would like to lend your expertise and serve in any of these specific groups please contact Wendy Wigen at .
●Representation group:
○To build upon a core triples model to add facilities for representing time, provenance, etc.
○Enable progress on prototypes, using them as testbeds for exploring different options for representations.
■Create a set of challenges/competitions to promote the construction of OKN. These will typically involve constructing a Knowledge Base, either automatically by reading text or from schema.org markup in web pages and then validating the KB with tests such as answering questions and/or populating Wikipedia at a given level of accuracy by Dec 2017.
■Release a major RFP for universities and other national centers of excellence to build components of a system with the following characteristics:
●Extractors to read the web and populate the OKN
●Data pipelines that normalize and integrate data from different sources into a single index, taking care of issues such as provenance, entity resolution, schema variations, etc.
●Testers to develop technologies to validate new entries into Wikidata using less than 1% of the human-level resources that are currently used but with increased precision at detecting good versus bad data.
●Knowledge representation research into how to extend the notion of probabilistic information, allowing OKN to represent uncertainty in a way which is practical and allows some useful form of uncertain reasoning and on incorporating provenance and other kinds of metadata into the representation. Internet search company commitment: We will attempt to have at least five major internet companies agree that subject to data quality guarantees, they will put incorporation of knowledge from the OKN into their products, at least in beta-testing form.
●Serving group:
○Develop a system in which users can upload, collaborate on and download data sets. The system should provide facilities for answering queries issued against specified datasets.
○Provision, via a suite of cloud providers, servers to host this system, so that it can be available as service from multiple providers. do matching and result-set returns for prototypes.
●Competition Group:
○Organize, design, and sponsor a competition that results in a prototype for OKN.
○Include the following characteristics:
■Specific enough to be useful and answer questions in a specific domain,
■Accessible enough to be used as a class project,
■Applicable to a compelling domain such as biomedical or education.
[1]Based on discussions at the Entities, Facts, Questions, Answers (EFQA) Meeting held on Friday July 29th, 2016, at the White House Office of Science and Technology Policy (OSTP), Eisenhower Executive Office Building, Washington, DC, sponsored by the Networking and Information Technology Research and Development (NITRD) Big Data Interagency Working Group. List of attendees is in the last section of the document.