NSF/Intel Partnership on Machine Learning for Wireless Networking Systems | 19 591

Opportunity Information: Apply for 19 591

The NSF/Intel Partnership on Machine Learning for Wireless Networking Systems (Funding Opportunity Number 19-591) is a discretionary grant program run by the National Science Foundation that aims to push forward fundamental research at the intersection of machine learning and wireless communications. The central idea is to develop wireless-specific ML methods that can reshape how future wireless networks are designed and operated, rather than simply applying generic ML tools to networking problems. The opportunity is positioned around the need for next-generation wireless systems to function well under tough real-world constraints: shared and crowded spectrum, limited radio and network resources, rapidly changing environments, and increasingly strict quality-of-service demands from emerging applications.

A major theme in the solicitation is that the program is not just looking for incremental performance tweaks. It is looking for research that can lead to new wireless system and architecture designs where ML plays a meaningful role in decision-making and adaptation. That includes networks that can dynamically access shared spectrum, intelligently allocate scarce resources, and scale to support diverse use cases with different latency, reliability, throughput, and fairness requirements. The program also emphasizes grounding the work in realistic wireless problems and explicitly encourages investigators to clarify when and why ML is the right tool, especially compared to traditional model-based approaches that have historically been very effective in building the wireless network stack used today.

In parallel to ML used inside the wireless network, the program also targets reliable distributed machine learning that operates over wireless edge networks. This second thrust focuses on the challenge of performing computation and learning when training or inference is spread across devices connected by wireless links. The solicitation highlights the difficulty of doing distributed ML under wireless constraints such as variable bandwidth, intermittent connectivity, device heterogeneity, energy limits, and noisy communication. Research in this area is intended to make distributed ML more robust and efficient in edge settings, enabling ML-driven wireless operation as well as supporting broader future applications that depend on learning at the edge.

Proposals can address one or more of three named Research Vectors (RVs): (1) ML for Wireless Networks, which broadly covers ML techniques for improving or rethinking wireless protocols and network operation; (2) ML for Spectrum Management, which includes learning-enabled approaches to sharing, sensing, allocating, and coordinating spectrum use in dynamic environments; and (3) Distributed ML over Wireless Edge Networks, which centers on methods for training and coordinating models across wireless-connected edge devices in a reliable and resource-aware way. Across these vectors, the solicitation signals interest in work that identifies foundational questions, establishes clear baselines against model-based methods, and explains what kinds of improvements are realistically achievable and under what assumptions.

In terms of funding, NSF anticipates making roughly 10 to 15 awards. The typical project duration can be up to three years. Award sizes are expected to range from about $300,000 to $1,500,000, with the budget expected to match the ambition and complexity of the proposed research. While an award ceiling is not specified in the listing, the program description indicates that projects will be funded within that stated range. The opportunity falls under NSF research and development activity categories and lists CFDA numbers 47.041 and 47.070.

Eligibility is listed broadly as "Others" with additional details referenced in the solicitation’s eligibility section, so prospective applicants would need to consult the full announcement text for precise eligibility rules. The opportunity was created on July 30, 2019, with an original closing date of October 29, 2019, indicating this specific call was time-bound; however, the program description provides a clear picture of the research direction NSF and Intel were aiming to catalyze: foundational, wireless-aware ML that can be justified against established modeling approaches, and distributed edge learning that is resilient to the realities of wireless communication.

• The National Science Foundation in the science and technology and other research and development sector is offering a public funding opportunity titled "NSF/Intel Partnership on Machine Learning for Wireless Networking Systems" and is now available to receive applicants.
• Interested and eligible applicants and submit their applications by referencing the CFDA number(s): 47.041, 47.070.
• This funding opportunity was created on Jul 30, 2019.
• Applicants must submit their applications by Oct 29, 2019. (Agency may still review applications by suitable applicants for the remaining/unused allocated funding in 2026.)
• The number of recipients for this funding is limited to 15 candidate(s).
• Eligible applicants include: Others (see text field entitled Additional Information on Eligibility for clarification).

Apply for 19 591

[Watch] Creating a grant proposal using the step-by-step wizard inside the applicant portal: