Research

Networking and AI are two of the most transformative IT technologies --- helping to better people’s lives, contributing to national economic competitiveness, national security, and national defense. The Institute will exploit the synergies between networking and AI to design the next generation of edge networks (6G and beyond) that are highly efficient, reliable, robust, and secure. A new distributed intelligence plane will be developed to ensure that these networks are self-healing, adaptive, and self-optimized. The future of AI is distributed because AI will increasingly be implemented across a diverse set of edge devices. These intelligent and adaptive networks will in turn unleash the power of collaboration to solve long-standing distributed AI challenges, making AI more efficient, interactive, and privacy preserving. The Institute will develop the key underlying technologies for distributed and networked intelligence to enable a host of future transformative applications such as intelligent transportation, remote healthcare, distributed robotics, and smart aerospace. 

Collaborations
AI-EDGE GLOBAL
Experimental Facilities
Research Products
Seminar Series
External Talks
two broad synergistic themes: AI for Networks and AI on Networks

Thrusts

The research plan of the Institute is organized around 8 thrusts that span two broad synergistic themes: AI for Networks (Thrusts 1-4) and AI on Networks (Thrusts 5-8).

AI for Networks

The astonishing success of AI provides a unique opportunity to design distributed intelligent efficient, self-healing, secure, and adaptive next generation edge networks. While  the  preliminary  successes  of  AI  for  networks  have  been  promising,  developing AI/ML algorithms to networking with minimal or no human oversight poses many important research questions will be explored systematically and in depth.

AI Chip and Wireless Tower
Re-engineering the Physics/Constraints
Re-engineer the physical fabric for6G+ wireless communications through AI, thus treating the fabric itself as a controllable entity.
AI Chip and Flow Chart
AI Based Network Resource Allocation
Develop new AI techniques for the design and control of these next generation networks taking into account practical resource constraints.
AI Chips with phones, cars, laptops, and drones, all interconnected
Multi-agent Network Control
Develop multi-agent AI techniques for distributed intelligence and control across possibly non-cooperative, network entities.
AI Chips and security locks
AI Powered Network Security
Develop new AI tools and techniques to guarantee that the network is secure, intrusion free, and highly robust.

AI on Networks

With the dramatic increases in processing power at the edge devices, we expect that the future of AI will be distributed as devices will use AI to make decision in their local environments connected by the network. We will develop intelligent and adaptive networks that will unleash the power of collaboration to solve long-standing distributed AI challenges, making AI more efficient, interactive, and privacy preserving.

AI Chips around a rotating gear
Network Aware AI Operation
Develop distributed AI that will seamlessly adapt its operation by taking into account computation, communication and data constraints.
AI Chips in a network with cars, cameras, drones, and robots.
Network Operations for Distributed AI
Re-engineer networks by adaptively allocating communication, computing, and storage resources for serving the needs of distributed AI applications.
Human and AI Chip interacting
Humans, AI and Network Research at the Interface
Develop new collaborative methods across humans-AI-networks to make systems more efficient than either human or machines by themselves.
AI Chip with a lock and a crossed out eye
Security and Privacy of Network Users
Design and control the networks such that they are privacy-aware and can be optimized to facilitate protection from information leakage and attacks.

Use Cases

Use Case 1
Use Case 2
Use Case 3

Use Inspired Research: The focus of the AI Institute is on edge networks, which will constitute most of the growth of future networks. This edge includes all devices connected through the radio access network as well as data centers and cloud computing systems that are not in the core backbone network. A critical component of the Institute is to shorten the duration between foundational research and use case research across multiple disciplines. This will result in a virtuous cycle that will have a cascading impact, dramatically accelerating the transition from research to implementation and technology transfer. Three wireless edge use cases have been developed to help connect research of the Institute to several practical applications

  1. Ubiquitous Sensing/Networking: This use case focuses on AI-driven sensing and networking. Testing of this use case will leverage the POWDER and COSMOS PAWR platforms being deployed in Salt Lake City and New York City over metropolitan areas of tens of miles in diameter. These networks will not merely carry information, but also generate, deliver, and process data from pervasively deployed multimodal sensors, enabling AI agents to become cognizant of the environment, necessary for optimized actuation within the wireless infrastructure.
  2. Human Machine Mobility: This use case focuses on extreme mobility settings. Testing for the use case will leverage a 6G+ system with terrestrial vehicular and unmanned aerial system (UAS) mounted systems being deployed at the AERPAW PAWR platform in Raleigh, NC the platform will be a distributed wireless infrastructure where mobile elements complement resource provisioning by fixed towers. By including mobility at various scales these systems can coordinate to create dynamic MIMO platforms to bridge the vast tracts of land that continue to impede rural broadband.
  3. Programmable/virtualized 6G networks: The focus here will be on open, agile, and modular radio access network architectures rooted in the principles of software-defined networking, virtualization, interoperability, and of separation of the data plane from control functionalities. We will demonstrate key outcomes in the Colloseum emulation platform, the world's largest RF emulator. The Colloseum allows instantiation of multiple virtual network slices that are tailored to accommodate diverse network services, tenants, and traffic on-demand.  This use case will lead to an unprecedented ability to control the entire network infrastructure end-to-end by AI.

These use cases connect the key research thrusts of the Institute to practical applications through testing on experimental platforms. The Institute will work with its industry and DoD partners to facilitate translation and adoption.

Synergy and Virtuous Cycles

The institute's research is built upon the synergistic relationships among the “AI for Networks” and “AI on Networks” research thrusts, with specific tasks that have inter-dependent challenges, which will bring together researchers from these different tasks, with progress on tasks in one thrust enabling or informing the other. It is precisely this web of inter-dependent research relationships and activities that make our Institute so much more than the sum of its parts.

Our institute's research forms a “virtuous cycle” (explicitly called out in NSF 20-503) that will be a core component of our efforts to ensure a tight coupling between foundational and use-inspired research. New algorithms, approaches, and system design/function emerging from the eight foundational research thrusts will be instantiated in three use case testbeds; experience and insights gained in these testbeds will inform further innovations in foundational research. An important second “virtuous cycle” exists between “AI for Networks” and “AI on Networks”. The foundational advances in AI – motivated and designed for, as well as constrained by, real-world system needs – will result in the enhanced underlying network and computational infrastructure for the AI algorithms.

In the virtuous cycle, use-inspired research informs the set of challenges (and constraints) undertaken by fundamental research. For example, AI-driven network control algorithms must operate in a network setting where organizational boundaries (e.g., backbone service providers, edge wireless networks, various sensor owner/operators, and fleet owners/operators) are important – data and state information may only be minimally-shared across such boundaries or only shared when there is economic gain (e.g., as in BGP policy-driven routing). This suggests that a new architecture will be needed, beyond the traditional Internet IP “hourglass”, for AI-driven networks. Driven by our use cases, the eight research trusts will help define and evolve this architecture and operate in the context of this architecture. They will identify new “thin waist(s)” where higher-level notions of data and information and lower-level notions of link abstraction, rather than ubiquitous end-end network connectivity alone, are likely to be critical. Using open-source software implementations we will maintain a shareable reference architecture and design incorporating the waists and support experiments with use cases as well as collaborative knowledge transfer to our partners.