Deep Green

DARPA (The Defense Advanced Research Projects Agency) is asking the IT world for proposals to create a state-of-the-art automated battlefield decision-making system. (50 page pdf)

The scene in Dr. Strangelove where Gen. Ripper and Group Captain Mandrake discuss the mid-20th century paradigm shift from the politicians to the generals vis à vis the capability of and responsibility for strategic military thought comes to mind.

The implication of the envisioned new system (called "Deep Green" by DARPA) is that it is time now for the generals to pass the baton to the computers.

The Defense Advanced Research Projects Agency (DARPA) Information Processing Technology Office (IPTO) seeks strong, responsive proposals from well-qualified sources for a new technology program called Deep Green. Deep Green will build a battle command decision support system that interleaves anticipatory planning with adaptive execution. Deep Green must be capable of addressing the full spectrum of joint and combined arms capabilities available to the modular brigade commander, drastically increasing the option and future space. This will allow the commander to think ahead, identify when a plan is going awry, and help develop alternatives "ahead of real time." The commander (and his support staff) is involved in essentially two major asynchronous functions: generating options and making decisions. The goal of this program is to create a commander-driven system to assist the commander and his support staff in generating options or Courses of Action (COAs).

Deep Green will aid in battle command and commander’s visualization by creating technologies that make it easier for the commander to articulate options to consider and anticipate the possible futures that result from those options. This proactive analysis will help predict which possible futures are becoming more likely – before they occur. Given that information, the commander can make better decisions and focus planning efforts (the generation of future branches and sequels) on where they can be the most useful. To accomplish this, Deep Green will focus on the following functional components: The Commander’s Associate (which consists of Sketch to Plan and Sketch to Decide), Crystal Ball, and Blitzkrieg.

There are six (6) tasks under the envisioned Deep Green program:

Task 1: Commander’s Associate (Sketch to Plan and Sketch to Decide)
Task 2: Blitzkrieg
Task 3: Crystal Ball
Task 4: Automated COA Generation
Task 5: Integration
Task 6: Test and Evaluation

(...)

The United States has a compelling need for reliable information affecting military command, soldiers in the field, and national security.

Today’s technical barriers include the following issues:

■ New technology is needed for machine induction of intuitively expressed plans.

■ Multi-modal (sketch and speech) collaborative technologies must be extended to incorporate modern learning technology that induces plans and the user’s intent from intuitive, coarse-grained plan descriptions.

■ Existing AI planning & monitoring systems

■ are largely deterministic in nature, while the battlefield is inherently stochastic;

■ focus on full automation rather than commander-driven plan generation; and

■ are reactive in nature, re-planning after the plan has broken

■ The current generation of combat models

■ run slowly,

■ generate a narrow spread of possible outcomes, and

■ require significant manual intervention.

The overall goal of Deep Green is to provide a technology that allows the commander
to:

■ generate and analyze options quickly, including generating the many possible futures that may result from a combination of friendly, enemy, and other coursesof action;

■ use information from the current operation to assess which futures are becoming more likely in order to focus the development of more branches and sequels; and

■ make decisions cognizant of the second- and third-order effects of those decisions.

Deep Green is composed of tools to help the commander rapidly generate courses of action (options) through multimodal sketch and speech recognition technologies. Deep Green will develop technologies to help the commander create courses of action (options), fill in details for the commander, evaluate the options, develop alternatives, and evaluate the impact of decisions on other parts of the plan. The permutations of these option sketches for all sides and forces are assembled and passed to a new kind of combat model which generates many qualitatively different possible futures. These possible futures are organized into a graph-like structure. The commander can explore the space of possible futures, conducting “what-if” drills and generating branch and sequel options. Deep Green will take information from the ongoing, current operation to estimate the likelihood that the various possible futures may occur. Using this information, Deep Green will prune futures that are becoming very improbable and ask the commander to generate options for futures that are becoming more likely. In this way, Deep Green will ensure that the commander rarely reaches a point in the operation at which he has no options. This will keep the enemy firmly inside our decision cycle.

(...)

Security classification guidance on a DD Form 254 (DoD Contract Security Classification Specification) will not be provided at this time since DARPA is soliciting ideas only and does not encourage classified proposals in response to this announcement. However, after reviewing incoming proposals, if a determination is made that contract award may result in access to classified information, a DD Form 254 will be issued upon contract award. If you choose to submit a classified proposal you must first receive the permission of the Original Classification Authority to use its information in replying to this announcement.