In October 2018, NSF implemented the Domain-based Message Authentication, Reporting & Conformance (DMARC) email changes required by the Department of Homeland Security (DHS) to improve email security. Some email routing practices (such as auto-forwarding to personal email accounts and sending messages through third-party providers) may cause messages to be flagged as potentially fraudulent by DMARC security checks and blocked. If your email is auto-forwarded to another account, such as a personal email account, you may not receive emails from NSF in that forwarded account. More information about DMARC and email delivery from NSF.
Division of Information and Intelligent Systems
Robust Intelligence (RI)
See program guidelines for contact information.
The Robust Intelligence program (RI) encompasses the broad spectrum of foundational computational research needed to understand and enable intelligent systems in complex, realistic contexts. Robust intelligence may be characterized by flexibility, resourcefulness, creativity, real-time responsiveness and long-term reflection; use of a variety of representation or reasoning approaches; ability to learn and adapt performance at human levels and beyond; and awareness of and competence in complex environments and social contexts.
Progress in this challenging endeavor requires advances across several computational disciplines that underpin the pursuit of intelligent systems. The RI program accepts research proposals aimed at contributing deeper understanding and new insights in artificial intelligence, computational neuroscience, computer vision, human language technologies, machine learning, and robotics. Researchers proposing foundational advances in RI research areas are encouraged to list the applicable discipline(s) first among keywords in the Project Summary, followed by descriptive keywords dealing with the specific methods or topics that are the focus of the proposal.
Proposals transcending RI disciplinary boundaries and fueling their intellectual co-evolution are highly encouraged. Such approaches might integrate methods from multiple computational disciplines to contribute foundational advances or other significant new knowledge and insights in robust intelligence. Advances may address progressively richer environments, larger-scale data and more diverse computing platforms, and more sophisticated computational and statistical approaches, looking to nature, in some cases, to model cognitive and computational processes. Researchers proposing advances through integrative approaches are encouraged to make this clear in the Project Summary and keywords. Proposals focused on the application and fielding of established methods are not reviewed in this program.
More information on topics of interest to the RI program is available at:
Funding Opportunities for the Robust Intelligence Program: