AFOSR DURIP HydraAI- LLM and GenAI Compute Platform to Support AFOSR Multi-Scale Hydra
Abstract
Complex networks are pervasive in critical infrastructures such as energy, transportation, and cyber systems. In order to operate successfully in uncertain and adversarial environments, complex networks must mitigate natural and malicious disruptions. The goal of network resilience is to develop mitigation strategies that ensure performance, connectivity, and functionality during and after a disturbance. Such mitigation strategies will need to incorporate both discrete (locations of nodes-edges) and continuous (control inputs) decision variables that affect network dynamics at multiple time-scales, creating a hybrid system. Our currently funded AFOSR Multi-Scale Hydra project plans to investigate and develop a dynamic hybrid submodular framework for network resilience across multiple time-scales. We will identify fundamental properties (e.g., safety, stability, connectivity) that need to be be maintained under topology changes, failures, and attacks, which we term resilient network invariants. To ensure computational tractability, we will explore hybrid submodular structures of the invariants, which generalize submodular optimization techniques from discrete functions to functions with both discrete and continuous variables. Our submodular techniques, in conjunction with deep neural networks (DNNs) will enable us to handle not only large scale network, but also the data that is generated by such large networks in order to ensure scalability and operational manageability. This AFOSR DURIP proposal titled HydraAI requests support to build a computing platform to harness transformative powers of emerging generative AI (GenAI) and large language models (LLMs) to support the research effort of our AFOSR Multi-Scale Hydra project. Having the HydraAI compute platform will enable the AFOSR Multi-Scale Hydra project to develop resilient solutions for existing and emerging complex networks. AFOSR Multi-Scale Hydra builds on our recently concluded project, AFOSR Hydra, which investigated efficient solutions for safe operation of resilient complex networked systems in adversarial environments. Our proposed HydraAI consists of a cluster of two platforms consisting of NVIDIA H100 GPUs, connected by a bridge. The platform will be critical to support research on the AFOSR Multi-Scale Hydra project to develop resilient future systems. The HydraAI platform will be housed in the Network Security Lab (NSL) in Room ECE 437 of the Electrical and Computer Engineering building at the University of Washington. UW ECE will provide the necessary climatecontrolled housing, power, networking, cooling, and acoustic dampening required for the smooth and sustained operation of the platform. The HydraAI platform will additionally impact our other ongoing projects supported by other DoD agencies including the ONR, and will contribute to education and workforce development.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Feb 06, 2025
- Source ID
- FA95502510043
Entities
People
- Linda Bushnell
Organizations
- Air Force Office of Scientific Research
- United States Air Force
- University of Washington