Realizing the Value of AI in Defense and Security

Defense and security organizations need rapid culture change in order to realize the value of their data and keep up with evolving technology.

The United States government is the single largest buyer in the world, spending over $500 billion each year, with a substantial percentage going into IT procurement. However, even with these resources, the U. S. government lags behind commercial organizations in terms of technical maturity.

With the Internet of Things (IoT) infiltrating the world of defense and security at an astounding pace, there is a vast amount of data available to intelligence and security agencies. However, collecting data is easy – using it efficiently is the challenge.

 

AI for Security and Defense: Main Challenges

  1. Organizing unstructured data from multiple sources. Unlike businesses in financial services, most of the valuable data in defense and security sector is unstructured: terrestrial, satellite-based data, static data about individuals or entities, agency databases, etc. Defense and security organizations use data from various sources, hence the challenge lies in combining, connecting, and correlating data stored in different databases with no central place to access everything easily. Data science, machine learning, and AI platforms can help alleviate this challenge.
  2. Dealing with sensitive data under time pressure. The most difficult challenge is extracting actionable insights in a scalable way and in the time frame needed for sensitive defense and security projects. Teams that can leverage a data science, machine learning, or AI platform will be more agile since every member of the team can explore data and leverage AutoML without relying only on data scientists. In addition, data science platforms allow processes and workflows to be easily reproduced; building blocks can be reused from old projects to seed new ones.
  3. Security and data governance: Good data governance and security practices become even more important when working with sensitive data for security and defense. Data science, machine learning, and AI platforms allow teams to monitor and control all modifications, work, and access to sensitive data. Centralized data efforts enable robust and secure governance of larger projects.
Watch Video

High-Value Use Cases

Crime prediction: Crime data has two specific dimensions – geographic and temporal (crimes happen in different places at different times). Additional data sets provide other information on the weather, neighborhood, and public transportation that can impact the final model. Not every point in time and space is equally likely to host a crime, but machine learning can leverage this data to help understand the factors that contribute to historical crime and even predict future crimes.

Threat alert systems: As the amount and severity of attacks grow, AI can provide real-time insights that screen frequent notifications and rank alerts based on their priority, enabling analysts to focus on the most important alert first.

Watch Video

Image recognition and deep learning: Dataiku took on a challenge organized by the NATO Innovation Hub, which was sponsored by the Allied Command Transformation. The challenge asked the participants to provide a solution for “Data Filtering and Fusing, Visualization, and Predictive Analytics” in an imaginary assistance mission scenario for a disease outbreak in a landlocked country, leading to a public health crisis complicated by the emergence of rebel groups attacking medical supplies. This is a great example of image recognition as a use case for defense – read the high-level summary or the in-depth technical version for more.

 

Dataiku for Security and Defense 

Dataiku is the platform democratizing access to data and enabling organizations to build their own path to AI.  Specifically, Dataiku can equip teams with the ability to:

  • Iterate models in production to drive rapid value creation. Models can be evaluated and tested in the complexity of real-world environments alongside analysts’  workflows to develop trust in the system.
  • Enable regulatory compliance and auditability. Dataiku offers robust user permission tracking, ensuring that only authorized users can access sensitive information. Additionally, version history and usage justification information ensure auditability in the event of a compliance audit. However, user permissions are only useful when they’re enforceable, and Dataiku supports organizational best practices surrounding secure data use and storage.
  • Securely process data and create machine learning models, with or without coding. Dataiku offers robust AutoML features for analysis by non-technical users (as well as to speed up technical users), cutting out inefficiencies and allowing for rapid iteration.

Responsible AI for a Sustainable Data Future

Dataiku is the basis of a responsible AI strategy, ensuring that models are accountable, architecture and infrastructure is sustainable, and data processes are governable.

Read more

Go Further

Opportunities for AI for Federal, Public Sector & Nonprofits

Traditionally, the public sector has lagged behind private industry when it comes to technological advancement, and data science has been no exception.

Learn More

Governance, Security, and Monitoring

Dataiku makes it easy for administrators to search and organize datasets as well as monitor access and user activity.

Learn More

Getting Started With Deep Learning

Deep learning's main advantage is that it can handle massive amounts of data - particularly unstructured - well. Getting started doesn't have to be hard by leveraging publicly available pre-trained deep learning models to begin.

Learn More

Making Open Source a Sustainable Part of AI Strategy

Open source technologies are critical, but they also have drawbacks that necessitate a larger infrastructure around the technology to make it work for the enterprise.

Learn More