Notice Announcing Research Opportunity Announcement for the Advancing Health Research through Ethical, Multimodal AI Initiative
Notice Number:
NOT-OD-24-095

Key Dates

Release Date:

April 15, 2024

Related Announcements

NOT-OD-24-114 - Notice of Change to Research Opportunity Title: Advancing Health Research through Ethical, Multimodal AI.

Issued by

Office of Data Science Strategy (ODSS)

Purpose

This Notice announces the publication of a new Research Opportunity Announcement (ROA) for the Advancing Health Research through Ethical, Multimodal Artificial Intelligence (AI) Initiative. The ROA was released on the NIH Office of Data Science Strategy website: https://datascience.nih.gov/MultimodalAI (See OTA-24-015). The Advancing Health Research through Ethical, Multimodal AI Initiative aims to develop ethically focused, data-driven, multimodal AI approaches to more closely model, interpret, and predict complex biological, behavioral, and health systems and enhance our understanding of health and the ability to detect and treat human diseases. 

Informational Webinar Date: April 19, 2024, 2:00-3:00 PM Eastern Time

Letter of Intent (LOI) Due Date (Optional): April 29, 2024, by 5:00 PM local time of applicant organization.

Proposal Due Date: May 16, 2024, by 5:00 PM local time of applicant organization. Late applications to this ROA will not be accepted.

Background

Multimodal AI (MAI) has emerged as a technique for analyses that can combine multiple types of data (e.g., text, images, videos) simultaneously, rather than relying solely on one modality (e.g., text or image), to generate outputs or predictions. New innovations in data fusion, model training, and model assessment and application are needed to fulfill the promise of MAI in biomedical and health research. New innovations are also needed to ensure ethical approaches and consideration for stakeholder needs drives the design at all stages of development: data selection and preparation, model training, model assessment and application. This initiative is an opportunity for biomedical researchers to collaborate with expert researchers in AI (computer science and applied mathematics) and ethics and to engage with stakeholders to ensure MAI capabilities meet their needs.

This initiative supports efforts to create ethics-driven and data-driven multimodal AI models for use in the biomedical, behavioral, and/or clinical fields. NIH envisions a portfolio of innovative projects that address systems level biomedical challenges using collaborative and participatory approaches for multimodal AI that will elucidate unique opportunities, risks, and challenges. Collaborative, co-design approaches should be used in which there is continual and iterative feedback between different components of the project (i.e. model development, data preparation and collection, etc.). Additionally, ethical considerations should be integral to all phases of development, as well as engagement with appropriate stakeholders. Projects are expected to inform considerations for the appropriate use of multimodal AI, relative to other methodologies, and take significant steps towards incorporation of ethical frameworks and co-design approaches in the multimodal AI lifecycle.

Funding Instrument

The funding instrument is the Other Transactions (OT) Award mechanism.   OT awards are not grants, cooperative agreements, or contracts, and use an Other Transactions Authority, provided by Congress. Policies and terms for individual OT awards may vary between awards. Each award is therefore issued with a specific Agreement, which is negotiated with the recipient and details specific terms and conditions for that award.

Award Information

OT mechanism allows for significant flexibilities to make adjustments needed to pursue catalytic and transformative initiatives. Award levels may increase or decrease over time based on programmatic needs, funding availability, and recipient performance. NIH expects to support up to 10 projects with annual budgets around $1-2M per project.

Initial Project duration is anticipated to be two years, but individual projects may be shortened or extended based on programmatic needs, funding availability, and recipient performance.

Inquiries

Please direct all inquiries to:

Agency Contacts: 

Scientific Research Contact(s): 

Laura Biven, Rui Pereira De Sa, Lori Scott-Sheldon  

Email: [email protected]

Financial / Agreements Officer

DJ Milliken

Email: [email protected]