(CLOSED) NSF 23-571: Building the Prototype Open Knowledge Network (Proto-OKN)

This program supports the creation of a prototype Open Knowledge Network — an interconnected network of knowledge graphs supporting a very broad range of application domains. Open access to shared information is essential for the development and evolution of artificial intelligence (AI) and AI-powered solutions needed to address the complex challenges facing the nation and the world. Knowledge graphs, which represent relationships among real-world entities, provide a powerful approach for organizing, representing, integrating, reusing, and accessing data from multiple structured and unstructured sources using ontologies and ontology alignment. Currently, private-sector investments in knowledge graphs power numerous consumer applications including web search, e-commerce, banking, drug discovery, advertising, etc. Undertaking a similar but inclusive, open, and community-driven effort and making use of publicly available data holds the potential to create a platform that would empower government and non-government users — fueling evidence-based policymaking, continued strong economic growth, game-changing scientific breakthroughs, while addressing complex societal challenges from climate change to social equity.

Projects funded by this program will provide an essential public-data infrastructure to power the next information revolution similar to the Internet — transforming our ability to unlock actionable insights from data by semantically linking information about related entities.

Proposals for research projects must fall into only one of the following themes, which should be clearly identified in the project title.

Theme 1 – Proto-OKN Use Cases

Proposals in this category will focus on the challenges posed by specific application(s)/use-case(s) and the relevant data sources for the respective use-case(s). Proposals should describe the software tools, data workflows, ontology, and knowledge design and implementation methods that would be employed to create the knowledge graph infrastructure from underlying data.

Key constituents including end-users and data providers should already have been identified and should be engaged in and committed to this effort. Proposals should explain how the overall information infrastructure, including the data, software, and knowledge graph structures will be maintained including roles and responsibilities of the various constituents engaged in the project. This solicitation especially seeks Theme 1 proposals that align with the priorities from the partnering agencies, as listed below in Section II.C.2.

Theme 1 proposals should clearly describe the data governance procedures that will be in place to ensure transparency and fair use of all the data. Proposals should also describe the essential data and AI ethics issues relevant to the use-case(s) and the ethics standards that will be used to guide the effort.

As mentioned, Theme 1 projects should be prepared to work in collaboration with Theme 2 projects for interoperable technical implementation at scale, following the timeline(s) of the Theme 2 award(s). Theme 1 projects should plan for their deliverables to be available and used by Theme 2 projects in demo, alpha, or beta mode during the duration of this 2-year effort as discussed in the Deliverables and Management Plan in subsection II.C.4 below. Also, Theme 1 projects should be prepared to work with the Theme 3 project to help prepare appropriate education/training materials. Proposals should identify the specific personnel who will be engaged in such collaborations with Theme 2 and Theme 3 projects, respectively.

Theme 2 – Proto-OKN Fabric

Proposals in this category must focus on development and prototyping of robust, scalable, cloud-based technical infrastructure for Proto-OKN based on leading-edge technologies. Theme 2 projects will be expected to support and demonstrate use of the prototype infrastructure by Theme 1 projects in Year 2 of the program. At that time, Theme 2 projects will need to deploy and run “production systems” and support Theme 1 users. Theme 2 projects are strongly encouraged to include industry partners as leads, co-PIs, or collaborators, or ensure that the project is able to adopt a production systems approach in their efforts.

Theme 2 proposals should clearly describe the proposed technical architecture and implementation considerations. Implementations are expected to be open and cloud-based, to enable future OKN efforts to continue to link to the overall Proto-OKN fabric. Adequate explanations should be provided in case the implementation has any closed/proprietary aspects, and in case it employs private cloud infrastructure.

Theme 2 proposals should be able to demonstrate:

• Technical skill, i.e., technical expertise and experience in implementing large-scale data projects and systems;
• Technical support capability, i.e., the ability and willingness to provide technical support to assist Theme 1 projects; and
• Open and collaborative approach, i.e., the experience and ability to develop open systems and the ability to collaborate with other teams as peers.

Theme 2 projects will provide the technical expertise needed to help link together Theme 1 data/knowledge graphs to create an overall linked Proto-OKN fabric. Theme 2 projects will plan for their infrastructure to be available for use by Theme 1 projects in alpha mode in 9 months and in beta mode in 18 months, as described in the Deliverables and Management Plan in subsection II.C.4 below. Collaboration between Theme 1 and Theme 2 projects should be on a peer-to-peer basis rather than one side dictating to the other, with adequate time and personnel effort dedicated for this interfacing effort. Theme 2 projects are also expected to collaborate with the Theme 3 effort to help develop the relevant educational and training materials.

This program seeks to fund up to two Theme 2 projects via this solicitation. The two projects will be required to merge their efforts together to create a single, common OKN Fabric to be deployed prior to the end of Year 2. Details of the merged effort will be worked out in Year 1 in collaboration with NSF. Subject to availability of funds, the merged OKN Fabric effort will be funded for an additional 2-year period in order to provide on-going support to Theme 1 efforts as well as any other OKN-related efforts from partner agencies.

Theme 3 – Proto-OKN Education and Public Engagement

A single award will be made in this category for creating educational and training materials and tools addressing the broad range of users of the Proto-OKN — from data providers to expert end users, technical developers, senior managers, non-expert users/citizens, and students at graduate, undergraduate, high school and middle school levels. For each distinct end-user community, the goal of the education and training materials is to help explain how the Proto-OKN impacts society and help prospective users understand how to engage. Proposals should describe how they would develop materials to help:

Explain the underlying concepts and techniques of the Proto-OKN and, more broadly, the concepts of knowledge representation and knowledge graphs and the importance of that approach to a broad range of end-users;

• Provide the conceptual information and functional details necessary for users to effectively use the system; and
• Provide the conceptual information and functional details necessary for content and technology contributors to effectively contribute to the system.
• As mentioned earlier, Theme 1 and Theme 2 projects will collaborate with the Theme 3 effort in order to help develop the relevant educational and training materials.

Proposals are encouraged to consider novel approaches and activities for effectively engaging with the broad range of end-users to facilitate effective user engagement.