Phase II of the Army's Deep Green Competition has Started

By Office of Business TransformationFebruary 23, 2023

(Photo Credit: U.S. Army) VIEW ORIGINAL

WASHINGTON – The Office of Business Transformation (OBT) has completed Phase I and is now entering Phase II of its annual Deep Green Data Science and Artificial Intelligence (AI) Competition.

Phase I produced remarkable results with all competitors outperforming the baseline model. Phase II will likely be no different.

Phase II of the competition will run from 9 January to 2 June, 2023, and focus on building a Semantic Segmentation model using raw images, point cloud data, or a fusion of both datasets. The goal of Semantic Segmentation is to classify every point/pixel on an image or LiDAR scan as a specific class.

Mr. Bakari Dale, the Army's Senior Advisor for Enterprise Data Science and Artificial Intelligence, wished all of the current Phase II participants good luck and emphasized the importance of what they are working towards: “I am proud of the work that we have done so far with the Deep Green Data Science & AI competition, and I am excited about the possibilities that lie ahead. My intent is that we use Deep Green to help prepare the workforce to leverage and operationalize AI, Machine Learning, and Deep Learning. The future digital workforce must know how to harness the power that can delivered by AI technology, and we cannot take our foot off the AI Accelerator pedal.  I believe that those who can control their data, and are leaders in AI, will shape the future of the world."

To spur further innovation in this phase of the competition, the Office of Business Transformation will offer a No-Code Deep Learning tool developed by This No-Code tool is intended to enable citizen data scientists across the DoD, with limited or no coding experience, the ability to participate in the competition and improve their understanding of Deep Learning and AI best practices.

Throughout the month of March, participants in the Deep Green Competition will also gain access to’s AI Cloud services, which will enable them to build Semantic Segmentation models with an easy-to-operate user interface.

For those who do not understand how to use, expert data scientists and researchers will provide guided training in how to use the tool and also provide the knowledge required to leverage HAIC.

For the more advanced data scientists, the HAIC platform will enable faster and more efficient prototyping and experiments.

According to Mr. Rohit Dhanda, VP Public Sector of, “ is delighted to support the Army Deep Green Data Science & Artificial Intelligence Competition. The H2O AI Cloud is available to participants who can now leverage Hydrogen Torch. A capability designed by top computer vision Kaggle Grandmasters. The H2O AI Cloud and Hydrogen Torch provide a low-code/no-code environment for semantic segmentation and computer vision.  Our goal is to ensure that all Challenge participants can quickly create fast and highly accurate models no matter their experience or skill level.”

The intent behind providing HAIC is to make the competition more accessible, and thereby grow the field of competitors. The inclusion of a low-code/no-code tool allows for coding skill diversity and gently introduces novice participants to a quickly growing sector of complex computational science.

According to the competition's manager, Dr. Matthew Millar, “The use of a no-code solution, like HAIC, is important to enable those with little to no experience in programming to compete in a very complex competition and learning environment. By offering a Graphical User Interface to interact with, users can see how each hyperparameter effects a model’s performance and can also easily learn the most useful principals of artificial intelligence. All without having to write code."

While the outcomes the participants produce are tremendously useful, Dr. Millar reminds us that this is not the only purpose of Deep Green. "One of the goals of Deep Green is to train DoD members in the best practices of Data Science, AI, and computing. A no-code tool lowers the bar for entry, allowing anyone the ability to participate and learn best practices in Deep Learning. The Deep Green initiative helps drive innovation, knowledge, and increases technology acceptance across all levels of leadership and domains.”

Registration for Phase II of the competition is still open. To register, please contact Dr. Matthew Millar at or use the Deep Green registration form at