How do we transform the DoD and accelerate AI adoption at scale? In a recent webinar, ViON’s Rob Davies posed that question to Colonel Sang Han, Chief of Infrastructure & Platform Division for the Joint Artificial Intelligence Center (JAIC), Kelli Furrer of Dell Technologies and Anthony Robbins from NVIDIA to discuss the future of AI for the Department of Defense.
Currently, the DoD is tackling IT modernization across all branches of the military, with AI representing a major push to aid decision-making and extract the greatest value from data. These are a few key insights industry leaders shared about their observations and experience in transforming IT to support AI within the DoD.
1. AI/Machine Learning Development is Being Done in Silos
Currently, infrastructure tools and data to build AI/ML applications is in disparate pieces and there is a growing need to develop an environment that can converge these capabilities. Having duplicative tools and trying to bring these tools into separate environments requires going through cybersecurity exceptions to onboard them, which equates to inefficiencies. The goal is to lower that barrier so that the developers have easier access to data, tools, and a common infrastructure, spread across departments.
What’s the path forward? Build a data catalog so that agencies can have data availability and accessibility to quickly build AI/ML applications models rather than doing duplicative work maintaining different environments across the DoD. The JAIC’s Joint Common Foundation (JCF) is a cloud-enabled AI platform that will accelerate the development, testing, and fielding of new AI capabilities and is making tools, models and code more accessible to everyone in DoD to build apps that will truly accelerate AI development.
2. Infrastructure Must-Haves: Processing at the Edge, On-Prem Capabilities & Software to Connect It
There are three critical infrastructure components to consider when preparing for AI. First, the work that is happening at the edge now and in the future will be really important, especially when faster networks can go from millions to billions and trillions of devices creating massive amount of data. Agencies need an improved ability to process data and train these neural networks at the edge. Secondly, agencies need on-prem capabilities in addition to public cloud. Just as all branches of the military have their own unique capabilities, one infrastructure will not and cannot serve all problems. Lastly, the software that connects all of these platforms and capabilities will be essential to ensure ease of operations, oversight and management.
3. AI Infrastructure Breaks Down Barriers
There are companies that specialize in serving every department within the federal government and in doing so have spent years identifying the common requirements that align to the needs of IT leaders. This public/private partnership has yielded solutions like AI infrastructure as-a-Service, which gives government access to the robust capabilities needed for mission success in a “ready to deploy” solution. Built with these common requirements in mind, AI Infrastructure as-a- Service makes modern technology accessible now. It enables organizations to test and scale. With the ease of a flexible financial model, solutions can be deployed in days not years – which can make all the difference in mission success.
The future of AI is unlimited. Proof of concepts are vital to testing, evaluating and implementing the most impactful outcomes for mission success! See the recording of the full webcast on Jumpstarting AI Using Infrastructure as-a-Service.