1. What is the Common Fund?

The NIH Common Fund, managed by the Office of Strategic Coordination in the Office of the Director, supports cross-cutting, trans-NIH programs that require participation by multiple Institutes and Centers. Common Fund programs are intended to be transformative, catalytic, synergistic, cross-cutting, and unique. This program is a short-term, goal-driven strategic investment, with deliverables intended to catalyze research across multiple biomedical research disciplines. More information can be found at https://commonfund.nih.gov/about. 
 

2. What makes a good application for a Common Fund program?

Common Fund programs are designed to support actively managed, milestone-driven projects that are different from mainstream studies such as R01 or P01. These projects are expected to work together to address key roadblocks in emerging fields that will transform the way biomedical research is conducted. Reviewers will evaluate whether a project proposes an incremental improvement or a significant advancement which will transform the field.
 

3. Why was the NIH Common Fund's Bridge to Artificial Intelligence (Bridge2AI) program created?

Rapid advancements in Artificial Intelligence and Machine Learning (AI/ML) analysis in recent years show promise in discovering new biomedical and behavioral research insights (that are often surprising). In 2018, members of the AI/ML and computational biomedicine research communities held a workshop, and in 2019 the AI Working Group of the Advisory Committee to the NIH Director developed a final report to propel progress of AI/ML methods in biomedical and behavioral research. These groups encouraged NIH to stimulate widespread adoption of AI/ML models that can tackle cross-cutting biomedical and behavioral research grand challenges. The NIH Common Fund’s Bridge2AI program was created in response to these recommendations and aims to harness the emerging power of AI/ML models for human health, beginning with an ‘analysis-first’ approach that values the machine-understandability of data and incorporates ethical principles surrounding data collection and use. 
 

4. What is the goal of the Bridge2AI program?

The goal of the Bridge2AI program is to generate flagship data sets and best practices for the collection and preparation of AI/ML-ready data to address biomedical and behavioral research grand challenges. This goal will be accomplished by:

• Generating new flagship biomedical and behavioral data sets that are ethically sourced, trustworthy, well-defined, and accessible
• Developing software and standards to unify data attributes across multiple data sources and across data types
• Creating automated tools to accelerate the creation of FAIR (Findable, Accessible, Interoperable, and Reusable) and ethically sourced data sets
• Providing resources to disseminate data, ethical principles, tools, and best practices
• Creating training materials and activities for workforce development that bridges the AI, biomedical, and behavioral research communities

5. How is the Bridge2AI program structured?

The Bridge2AI program will use biomedical and behavioral research grand challenges to drive the development of ethics, standards, tools, data sets, and skills and workforce development strategies for linking scientific workflows, protocols, and other information about the data collection process into computable knowledge. The Bridge2AI program plans to support interdisciplinary Data Generation Projects (OTA-21-008) and a complementary cross-cutting Integration, Dissemination, and Evaluation (BRIDGE) Center (RFA-RM-21-023). Teams funded through these two planned opportunities will be expected to interact and collaborate regularly to complete cross-cutting Bridge2AI program goals.
 

6. Will Bridge2AI help make existing data AI-ready?

No, the Bridge2AI program is focused on generating new AI/ML-ready data that can be used to address biomedical and behavioral research grand challenges. In a separate effort, the Office of Data Science Strategy has an initiative to improve the AI-readiness of existing, IC-supported data that aligns with the NIH Strategic Plan for Data Science.
 

7. What is the Common Fund Data Ecosystem (CFDE) and will the Bridge2AI program participate?

The Common Fund Data Ecosystem (CFDE) is intended to make Common Fund data sets more impactful by enabling data discovery, access, interoperability, and reuse. Common Fund data sets include large data sets and associated digital tools needed to mine and analyze the data. To help maximize their impact, CFDE is creating a portal or central access point for Common Fund data sets, tools, and other digital objects, through which users can access and compute on data in a cloud environment. This CFDE portal is not designed as a data repository or platform but will instead link multiple data platforms that have been established through Common Fund programs and create cloud workspaces for users to access and compute on data across the different platforms. While users may continue to access an individual data set via the platform created for that data set, the CFDE will foster new discoveries and support different types of analyses by enabling queries of multiple data sets simultaneously. The CFDE will provide user support through automated help tools, online courses, webinars, and in-person training events. By providing infrastructure, advice, resources, best practices, common components, and services to fit the needs of Common Fund programs, the CFDE will increase the utility of Common Fund digital objects by making them more findable, accessible, interoperable, and reusable, also known as FAIR. The CFDE will require a collaborative effort, involving Common Fund Data Coordinating Centers who represent the data generators for each program, and a central coordinating component known as the CFDE Coordinating Center (CFDE-CC).

The Bridge2AI program will be structured to facilitate engagement in the CFDE. Bridge2AI projects are expected to interact with the CFDE to ensure the interoperability and compatibility of common software standards and tools with the CFDE and other Common Fund datasets and resources. Interoperability and compatibility with the CFDE includes, but is not limited to: sharing persistent identifiers of publicly available data sets, submitting metadata harmonized to a CFDE data model, and ensuring access to Bridge2AI data assets in a cloud environment.

For more information, please visit https://commonfund.nih.gov/dataecosystem/faqs.
 

8. What is the Science and Technology Research Infrastructure for Discovery, Experimentation, and Sustainability (STRIDES) Initiative?

As the scale of data-related research grows, NIH continues to explore opportunities to support NIH researchers’ use of the cloud for data storage and compute. The STRIDES Initiative enables NIH and NIH-funded investigators to more easily take advantage of cloud services. The STRIDES Initiative has negotiated agreements with commercial cloud service providers (CSPs) for 1) discounted rates on cloud services, 2) access to professional services to provide support on how to use the cloud effectively, 3) access to training to help learn more about the cloud and its capabilities, and 4) the potential for collaborative engagements to explore topics of special interest to NIH.

For more information, please visit https://commonfund.nih.gov/dataecosystem/faqs.
 

9. Where can I find more information about Bridge2AI?

Stay tuned for updates on funding announcements and the latest news from the Bridge2AI program by visiting the program website: www.commonfund.nih.gov/bridge2ai regularly and subscribing to the Bridge2AI listserv.
 

10. Who should I contact with questions?

Questions about the Bridge2AI program can be sent to [email protected].

11. What is Artificial Intelligence (AI) and is it different than Machine Learning (ML)?

Artificial Intelligence (AI) is a broad concept that encompasses the idea of machines being able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages. Machine Learning (ML) is an application (or subset) of Artificial Intelligence (AI) that provides machines (like computers) the ability to automatically learn and improve from experience, without being explicitly programmed. The intention of ML is to enable machines to learn by themselves using provided information (data) to make accurate predictions.
 

12. Why is AI/ML being used for research? How will AI/ML help improve human health?

There are a number of biomedical and behavioral research applications that would benefit from AI/ML approaches. For example, using AI/ML to decode the functional genome in order to understand how genomic variations affect outcomes in health, or to develop digital twins—a virtual representation of a living system used to monitor health, predict adverse events, assist clinical decision making, and/or personalize treatments. Using AI/ML approaches to help computers “understand” the conditions in which data are collected and the questions that data are intended to answer will unlock the potential of this information to transform science and human health.

General FAQs

13. What will the Bridge2AI Integration, Dissemination, and Evaluation (BRIDGE) Center do?

The BRIDGE Center will be responsible for integrating activities and knowledge across Data Generation Projects, disseminating products, best-practices, and workforce development materials/activities, and evaluating all aspects of the Bridge2AI program with input from external stakeholder communities. The BRIDGE Center will combine cores of expertise in team science, ethics, standards, tools optimization, and skills and workforce development. Please see RFA-RM-21-023 for additional information.
 

14. How many BRIDGE Centers will the Bridge2AI program support?

The Bridge2AI program plans to support one BRIDGE Center that will include multiple cores:

• Administrative Core
• Teaming Core
• Ethics Core
• Standards Core
• Tool Optimization and Dissemination Core
• Skills and Workforce Development Core

15. When will BRIDGE Center Team Building Activities take place?

A series of Bridge2AI Team Building Activities for potential applicants took place in June 2021. Visit the Bridge2AI Team Building Platform (registration required) for access to resources for building cross-disciplinary teams. Applicants may submit proposals to RFA-RM-21-023 if they did not participate in the Team Building activities.

For additional information visit the Scientific Meetings webpage.
 

Application FAQs

16. Should a potential BRIDGE Center application include multiple cores?

Yes, each potential BRIDGE Center application should include an Administrative Core and one or more of the following optional cores: Teaming, Ethics, Standards, Tool Optimization, Skills and Workforce Development. Cores will be individually reviewed and individual cores from different applications may be selected and funded to work collaboratively as a single BRIDGE Center. Please see RFA-RM-21-023 for additional information.
 

17. How should infrastructure activities be included in BRIDGE Center proposals?

The proposal for the U54 (RFA-RM-21-023) is a means to share with us your strategy, knowledge and experience - what you can bring to fulfill the spirit and goals of the Bridge2AI program. Please note the responsiveness criteria in the FOA guidelines: RFA-RM-21-023. The final determination of actual infrastructural needs will be determined by the Bridge2AI Consortium.

General FAQs

18. What is a Data Generation Project "grand challenge" and how do I know if my idea fits the criteria? <

Bridge2AI Data Generation Projects will be centered around biomedical and/or behavioral research "grand challenges" that promote broad uses for modern AI/ML models. Grand challenges should tackle complex biomedical questions that cannot be answered without the use of AI/ML approaches. When evaluating a Bridge2AI grand challenge idea, consider these three questions (1) have the AI/ML-ready data that could allow the primary question to be answered already been created? (2) can this question be answered effectively using existing analytical tools that do not use AI/ML? and (3) would the project fit within the general scale of a large R01 project? Bridge2AI grand challenges will likely have "no" as the answer to all three of these questions. Examples of grand challenges are provided in OTA-21-008, additional information about grand challenges can be found on the Bridge2AI IMAG Wiki page.
 

19. What is an Other Transactions (OT) Award?

The Other Transactions Authority is governed by 42 U.S. Code § 282 (n)(1)(b). Other Transactions (OT) are not grants, cooperative agreements, or contracts. They are used by the NIH to provide considerable flexibility in establishing policies for the awards; policies and terms for individual OT awards may vary between awards, each negotiated with a specific agreement, which may be expanded, modified, partnered, not supported, or later discontinued based on program needs, changing research landscape and or availability of funds.
 

20. Who is eligible to apply for an Other Transaction (OT) Award?

Prospective applicants may or may not have received NIH funding in the past. All entities public and private, small or large, for-profit or not-for-profit, eligible Agencies of the Federal Government (including NIH Intramural Research Program) are eligible to apply. Unaffiliated individuals who are U.S. citizens are also eligible to apply. For a list of anticipated eligible organizations please see the Other Transaction Opportunity Announcement (OTA-21-008).
 

21. How do I form a team around my Grand Challenge Team Building Idea?

Visit the Bridge2AI Team Building Platform (registration required) for access to resources for building diverse and cross-disciplinary teams. Applicants may submit proposals to OTA-21-008 if they did not participate in the Team Building activities.

For additional information visit the Scientific Meetings webpage.
 

22. When will Data Generation Projects Team Building Activities take place?

A series of Bridge2AI Team Building Activities for potential applicants took place in June 2021. Visit the Bridge2AI Team Building Platform (registration required) for access to resources for building cross-disciplinary teams. Applicants may submit proposals to OTA-21-008 if they did not participate in the Team Building activities. For additional information visit the Scientific Meetings webpage.
 

Application FAQs

23. For Data Generation Project OT applications, are the same documents that typically go in a traditional grant application required?

No. Other Transactions (OT) do not require the same documentation as traditional NIH Grants. Only the documents listed in OTA-21-008 are required to be submitted with the application. These include a Cover Page and a separate Research Plan and Budget for each Data Generation Project Module, a Resource Sharing Plan, a letter of support from the proposer’s organization indicating institutional commitment to the project, and letters of support from proposed collaborators who are not included as Key Personnel.
 

24. What should be included in the 2-page biosketch?

The 2-page biosketch should include pertinent information about the individual that would aid the reviewer in understanding that person’s qualifications and expertise pertaining to the project. The Bridge2AI Research Opportunity Announcement (OTA-21-008) indicates that, at a minimum, the information in the biosketch should include the name and position title, education/training (including institution, degree, date (or expected date), and field; list of positions and employment in chronological order (including dates); Other Research Support; and a personal statement that briefly describes the individual’s role in the project and why they are well-suited for this role. The NIH biosketch format used for grant applications is acceptable, but not required, as OT awards are not grants.
 

25. How do I submit a budget with my Data Generation Project OT application?

Applicants must complete a SF424 budget. Do not complete the budget form in the ASSIST module, instead download and complete the relevant form(s) found on the Bridge2AI Data Generation Projects Budget Forms and Instructions webpage. The completed SF424 budget PDF form(s) will then need to be included in the application as an “other attachment” in ASSIST. Budget information and any related administrative documentation shall not count toward the total proposal page limit.
 

26. Is one institution the prime (submitting institution) for all modules?

Yes. The Bridge2AI program requires each Data Generation Project to identify a Principal Investigator/Project Director (PI/PD) team with at least one biomedical/behavioral domain expert and one data science/data management expert, who are responsible for the overall leadership of the Data Generation Project. This team may be drawn from the leadership of any of the six modules. The application may be submitted by any member of this team. That member's institution will then be the prime for all modules.

Applicants should prepare editable budget spreadsheets using the form(s) available on the Bridge2AI Data Generation Projects Budget Forms and Instructions webpage. Please note that post-review, NIH will conduct negotiations with proposers to determine optimal scientific, technical, and logistical arrangement of each Data Generation Project award and their modules.

27. Should applications to each module have a separate budget and/or multiple subcontract sites?

Yes, each module must have a separate budget. Multiple subcontract sites are not required, but may be included as long as a justification is provided. The Data Generation Project personnel and budget structure should be set up based on the needs of the project. Applicants should prepare editable budget spreadsheets using the form(s) available on the Bridge2AI Data Generation Projects Budget Forms and Instructions webpage. Please note that post-review, NIH will conduct negotiations with proposers to determine optimal scientific, technical, and logistical arrangement of each Data Generation Project award and their modules.
 

28. Can participants in an ongoing project (outside of Bridge2AI) collect new multimodal data that do not fall within the original data collection/sharing plans as new data for Bridge2AI?

Yes, this could be new data for Bridge2AI, but with the caveat that the existing participants (and researchers) already represent a variety of perspectives. 
 

29. Can Bridge2AI Data Generation Projects include the development of AI/ML models?

No, Bridge2AI Data Generation Projects will not support the development of AI/ML models, they are intended to generate new biomedical and behavioral datasets that are ethically sourced, trustworthy, well-defined, and accessible. The purpose of the program is to create a bridge to AI by making the data AI/ML-ready, not by developing AI/ML models. Existing AI/ML models can be used throughout the data generation process in preparing the data, and mechanistic modeling frameworks can be used to determine what data need to be collected within the landscape of knowledge. Mathematical, statistical, computational (ML) models can be used as tools to incorporate ethical principles and standards into the data acquisition protocols, from multiple perspectives. After the Bridge2AI Consortium is formed, modelers will be encouraged to join in community activities (such as jamborees and codeathons) to mine and assess the Bridge2AI data for quality and ability to address grand challenges for future scientific discovery.
 

30. Should a budget for cloud compute and storage be included in a Bridge2AI application?

Yes, applicants should include any required cloud compute and storage costs in their application. If recommended for funding, NIH will consider these cloud costs and the possibility of leveraging the NIH STRIDES initiative for those costs. See FAQ #7 (above) for additional information about the STRIDES Initiative.
 

31. Should infrastructure needs, shared Institutional Review Boards (IRBs), and/or data sharing agreements be included in Data Generation Project module proposals?

A proposal for the Data Generation Projects (OTA-21-008) is a means to share with us your strategy, knowledge, and experience - what each applicant can bring to the table to help fulfill the spirit and goals of the Bridge2AI program. If including infrastructure needs, such as the amount of cloud resources needed to store a certain type of data or data sharing agreements needed to share and assess data, depicts your understanding of the challenges based on past experience, then propose it! The final determination of actual infrastructural needs will be determined by the Bridge2AI Consortium.
 

32. Does the bibliography count towards the Research Plan 10-page limit for each module?

No, the bibliography does not count towards the 10-page limit for each module's Research Plan.
 

33. In the Module Cover Page instructions for OTA-21-008, who should be listed as the "Awardee"?

In the Module Cover Page instructions on page 19 of OTA-21-008, the term “Awardee” refers to the Institution of the Contact Principal Investigator for that module (i.e., the Module Lead). The Module Cover Page should include the following information about the Module Lead and their Institution:

• Part C should provide the legal entity name, address, and contact information of the Module Lead’s Institution
• Part D should provide the name and contact information for the Module Lead
• Part E should provide the name and contact information for the Authorized Business Official of the Module Lead’s institution.
   

34. When will I hear back about my application?

Due to the unanticipated number of applications received, the review timeline posted in the Research Opportunity Announcement (OTA-21-008) is delayed. NIH staff are currently considering reviewer comments and OT award negotiations are expected to take place through February 2022.