How Partnerships Help Agencies Accelerate Custom AI Deployments

Artificial intelligence (AI) and machine learning (ML) are poised to transform military and defense decision-making and operations. In fact, according to a recent congressional research report, AI research is already well underway. There are several factors driving this transformation; namely, the availability of “big data” sources, improvements to machine learning approaches, combined with increases in compute processing power. 

Deploying AI models at the edge

For the first time, federal agencies can integrate AI and ML in ways and locations never possible before. One, for example, is directly at the network edge through high performance compute (HPC) systems, housed in rugged, server-class machines, featuring one or more graphics processing units (GPU) for accelerated processing. When the data at the edge is fed into new technologies that enable AI, agencies can make faster and more accurate decisions in the field, on the front line of every mission.

From autonomous vehicles to cybersecurity

What do AI applications in military and defense look like in practice? One obvious example is the use of AI technologies in autonomous and semi-autonomous vehicles. According to the Artificial Intelligence and National Security Report of November 10, 2020, applications in this field use AI technologies to perceive the environment, recognize obstacles, plan navigation, and even communicate with other vehicles.

Cybersecurity is another key area. In fact, as referenced  in his 2016 testimony before the Senate Armed Services Committee, Commander of U.S. Cyber Command Admiral Michael Rogers stated, “If you can’t get some level of AI or machine learning with the volume of activity you’re trying to understand when you’re defending networks … you are always behind the power curve.” As traditional cybersecurity tools look for historical matches to known malicious code, hackers need only change small sections of that code to break the defense. In contrast, Machine Learning can be used to identify suspicious network activities, potentially leading to earlier detection of cyberattacks, while AI-enabled tools can be trained to identify anomalies in broader patterns of network activity with federal agencies benefiting from much higher levels of protection.

Predictive maintenance powered by AI

Meanwhile, within the Department of Defense, the Innovation Unit (DIU) and the U.S. Air Force are working with the Artificial Intelligence Center of Excellence (JAIC) to produce prototypes of Predictive Maintenance solutions. According to the 2018 Department of Defense AI Strategy Summary, these commercially developed AI-based applications have the potential to predict equipment maintenance needs more accurately. Anticipated outcomes include improved equipment availability and reduced costs. Similar technology will be deployed to guide provisioning of spare parts and optimize inventory levels.

Customized hardware for tough and tight locations

However, deploying AI models at the edge is not without challenge. Apart from limited AI skills, many federal environments are too unique for off-the-shelf solutions but still require the security, reliability, and support of a Tier 1 infrastructure provider. Typically, they need customized hardware, capable of operating in tough environments with extreme temperatures, excessive dust, and high vibration levels. Space is often limited with solutions needing to fit into tight corners in temporary field offices, airplanes, ships or even the back of Humvees.

Security and performance demands

And of course, quality AI needs constant feedback loops and model retraining, requiring the need to federate data and update models over the air with minimal downtime. Apart from enterprise-level security, there are also performance demands. As AI and ML solutions need high processing power for data analysis, an increasing number of federal agencies and system integrators now look for GPUs to be part of the baseline configuration.

The right partner matters

Dell Technologies OEM Solutions has a long and successful track record of developing solutions for Federal customers that can capture, process, store, manage, and importantly, protect data and the insights generated in these rugged, non-traditional environments. We have a dedicated engineering team that can tailor and customize products to meet unique customer needs. Our long-life solutions make customer transitions easier and less disruptive, helping federal agencies maximize resources and budgets.

GPU-accelerated infrastructure

And importantly, thanks to our collaboration with NVIDIA, Dell Technologies offers GPU-accelerated infrastructure, pre-validated baseline software development toolkits, software containers, and reference architectures. Compared to traditional CPU architectures, these accelerators perform parallel computation and very fast execution of AI decisions. Looking ahead, I believe that these technologies will play an increasingly bigger role in managing AI and ML workloads. Yet another benefit of our work with NVIDIA is that we can help build out AI proficiency within federal agencies.

From threat detection to postal delivery

Let me share two examples of how we’ve partnered to develop turnkey, customized AI applications for Federal customers. Digital Harmonic offers a software AI application, PurePixel, which provides programmatic image enhancement in support of threat detection and mitigation. The solution is powered by pre-validated and configured PowerEdge servers, using NVIDIA GPUs and optimized software from NVIDIA NGC, a hub of GPU-optimized HPC and AI software.

We’ve also worked with Kinetica to develop a pre-configured solution to support its leading analytical in-memory databases. Designed on PowerEdge servers and leveraging NVIDIA GPUs, Kinetica can take large, complex streaming data sets – both structured and unstructured – and run analytics at the edge, including location-based and predictive analytics. Using visualizations and analytics of real-time data, the U.S. Postal Service has successfully deployed the Kinetica solution to efficiently deliver goods to more than 154 million addresses across the U.S. Both inspiring stories are told in this video above. 

Bring AI to life

Realize the potential of AI for your mission now! Dell Technologies OEM Solutions and its key partners will work with you to design, test, validate and deploy solutions that provide AI capabilities, while meeting all the rigorous operating conditions associated with federal environments.

  • Learn more about Partnerships for Custom AI Deployments:
  • Learn more about what we do for Federal customers.
  • Stay in touch. Follow us on Twitter @DellTechOEM and join our LinkedIn OEM Solutions Showcase page.



Source link

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.