The DARPA High-Performance Knowledge Bases Project

notion that includes not only the PQ53 [During/After ,] what {risks, rewards} would face in ? = {[exposure of its] {supporting, sponsoring} in , successful terrorist attacks against ’s , (PQ51), taking hostage citizens of , attacking targets with } = {terrorist group, dissident group, political party, humanitarian organization} Articles WINTER 1998 29 Answer(s): 1. Economic sanctions from {Saudi Arabia, GCC, U.S., U.N.} • The closure of the Strait of Hormuz would violate an international norm promoting freedom of the seas and would jeopardize the interests of many states. • In response, states might act unilaterally or jointly to impose economic sanctions on Iran to compel it to reopen the strait. • The United Nations Security Council might authorize economic sanctions against Iran. 2. Limited military response from {Saudi Arabia, GCC, U.S., others}... Source(s): • The Convention on the Law of the Sea. • (B5) States may act unilaterally or collectively to isolate and/or punish a group or state that violates international norms. Unilateral and collective action can involve a wide range of mechanisms, such as intelligence collection, military retaliation, economic sanction, and diplomatic censure/isolation. Figure 2. Part of the Answer Key for Question 53. Figure 3. A Parameterized Question Suitable for Generating Sample Questions and Test Questions. form and an order of battle that describes the structure and composition of the enemy forces in the scenario region. Given these input, movement analysis comprises the following tasks: First is to distinguish military from nonmilitary traffic. Almost all military traffic travels in convoys, which makes this task fairly straightforward except for very small convoys of two or three vehicles. Second is to identify the sites between which military convoys travel, determine which of these sites are militarily significant, and determine the types of each militarily significant site. Site types include battle positions, command posts, support areas, airdefense sites, artillery sites, and assembly-staging areas. Third is to identify which units (or parts of units) in the enemy order of battle are participating in each military convoy. Fourth is to determine the purpose of each convoy movement. Purposes include reconnaissance, movement of an entire unit toward a battle position, activities by command elements, and support activities. Fifth is to infer the exact types of the vehicles that make up each convoy. About 20 types of military vehicle are distinguished in the enemy order of battle, all of which show up in the scenario data. To help the technology base and the integration teams develop their systems, a portion of the simulation data was released in advance of the evaluation phase, accompanied by an answer key that supplied model answers for each of the inference tasks listed previously. Movement analysis is currently carried out manually by human intelligence analysts, who appear to rely on models of enemy behavior at several levels of abstraction. These include models of how different sites or convoys are structured for different purposes and models of military systems such as logistics (supply and resupply). For example, in a logistics model, one might find the following fragment: “Each echelon in a military organization is responsible for resupplying its subordinate echelons. Each echelon, from battalion on up, has a designated area for storing supplies. Supplies are provided by higher echelons and transshipped to lower echelons at these areas.” Model fragments such as these are thought to constitute the knowledge of intelligence analysts and, thus, should be the content of HPKB movement-analysis systems. Some such knowledge was elicited from military intelligence analysts during programwide meetings. These same analysts also scripted the simulation scenario. physical geography of a conflict but also the plans, goals, and activities of all combatants prior to, and during, a battle and during the activities leading to the battle. Three battlespace programs within DARPA were identified as potential users of HPKB technologies: (1) the dynamic multiinformation fusion program, (2) the dynamic database program, and (3) the joint forces air-component commander (JFACC) program. Two battlespace challenge problems have been developed. The Movement-Analysis Challenge Problem The movement-analysis challenge problem concerns high-level analysis of idealized sensor data, particularly the airborne JSTARS moving target indicator radar. This Doppler radar can generate vast quantities of information—one reading every minute for each vehicle in motion within a 10,000square-mile area.2 The movement-analysis scenario involves an enemy mobilizing a full division of ground forces—roughly 200 military units and 2000 vehicles—to defend against a possible attack. A simulation of the vehicle movements of this division was developed, the output of which includes reports of the positions of all the vehicles in the division at 1minute intervals over a 4-day period for 18 hours each day. These military vehicle movements were then interspersed with plausible civilian traffic to add the problem of distinguishing military from nonmilitary traffic. The movement-analysis task is to monitor the movements of the enemy to detect and identify types of military site and convoy. Because HPKB is not concerned with signal processing, the input are not real JSTARS data but are instead generated by a simulator and preprocessed into vehicle tracks. There is no uncertainty in vehicle location and no radar shadowing, and each vehicle is always accurately identified by a unique bumper number. However, vehicle tracks do not precisely identify vehicle type but instead define each vehicle as either light wheeled, heavy wheeled, or tracked. Low-speed and stationary vehicles are not reported. Vehicle-track data are supplemented by small quantities of high-value intelligence data, including accurate identification of a few key enemy sites, electronic intelligence reports of locations and times at which an enemy radar is turned on, communications intelligence reports that summarize information obtained by monitoring enemy communications, and human intelligence reports that provide detailed information about the numbers and types of vehicle passing a given location. Other input include a detailed road network in electronic The second challengeproblem domain for HPKB is battlespace reasoning. Battlespace is an abstract notion that includes not only the physical geography of a conflict but also the plans, goals, and activities of all combatants prior to, and during, a battle and during the activities leading to the battle. Articles