Federal agency name: defense advanced research projects agency (DARPA), defense sciences office (DSO) funding opportunity title: young faculty award (YFA) funding opportunity number:…

The Defense Advanced Research Projects Agency (DARPA), Defense Sciences Office awarded a $500,000 cooperative agreement to Northeastern University on September 30, 2025, under the Research and Technology Development program (CFDA 12.910).

The recipient will design algorithms to extract and incorporate domain-specific knowledge from foundation models into low-size, weight, and power (SWAP) deep neural networks for tactical mobile devices deployed in warfighting scenarios. The work addresses real-time situational awareness while protecting against adversarial machine learning and domain shifts. Northeastern will research real-time out-of-distribution detection algorithms to recognize and remove intentional perturbations and develop real-time knowledge editing algorithms that surgically modify deployed DNNs to remove low-priority domains and incorporate newly encountered domains. The proposed techniques will be evaluated against the state of the art with low-SWAP embedded devices.

The period of performance runs from September 30, 2025, through September 29, 2027. Work is performed in Massachusetts. The award is structured as a cooperative agreement, which signals anticipated involvement or collaboration between DARPA and the recipient beyond the typical grant relationship. This Young Faculty Award (YFA) funds revolutionary advances in science, devices, or systems rather than evolutionary improvements to existing practice.

Federal agency name: defense advanced research projects agency (DARPA), defense sciences office (DSO) funding opportunity title: young faculty award (YFA) funding opportunity number: DARPARA2501CATALOG of federal domestic assistance (cfda) number(s): 12.910 research and technology developmentprogram description: DARPA is soliciting innovative research proposals in the areas of interest to the following DARPA technical offices: defense sciences office (DSO), information innovation office (I2O), microsystems technology office (MTO), and the tactical technology office (TTO). Further detail regarding the specific technical areas of interest can be found below under open office topics. Proposed research should investigate innovative approaches that enable revolutionary advances in science, devices, or systems. Specifically excluded is research that primarily results in evolutionary improvements to the existing state of practice.executive summary. Tactical mobile devices with low size, weight, and power (swap) deployedin warfighting scenarios increasingly require the execution of deep neural networks (dnns) toachieve real-time situational awareness while being protected from adversarial machine learningand domain shifts. To proactively incorporate domains that the mobile device is expected to encounterduring its mission, we will design algorithms to extract and incorporate domain-specificknowledge from foundation models into a domain-agnostic low-swap DNN. After deployment, toaddress domains that were not accounted for before deployment, we will research real-time out-ofdistribution(OOD) detection algorithms to recognize and remove intentional perturbations, as wellas real-time knowledge editing algorithms that will surgically modify the low-swap DNN to removelow-priority domains and include the newly encountered domain. The proposed techniqueswill be evaluated against the state of the art with low-swap embedded devices.