CS 785, Fall 2001 Gheorghe Tecuci tecuci@cs.gmu.edu http://lalab.gmu.edu/ Learning Agents Laboratory Department of Computer Science George Mason University G.Tecuci, Learning Agents Laboratory Sample questions Define the problem reduction approach to problem solving. What is an instance? What is a concept? What is a positive example of a concept? What is a negative example of a concept? Give an intuitive definition of generalization. What does it mean for concept A to be more general than concept B? Indicate a simple way to prove that a concept is not more general than another concept. Given two concepts C1 and C2, from a generalization point of view, what are all the different possible relations between them? What are the basic elements in the definition of a property or a relation? Briefly define a plausible version space rule. G.Tecuci, Learning Agents Laboratory Sample questions What is a generalization rule? What is a specialization rule? What is a reformulation rule? Name all the generalization rules you know. Briefly describe and illustrate with an example the “turning constants into variables” generalization rule. Define and illustrate the dropping conditions generalization rule. Define the following: • a generalization of two concepts • a minimally general generalization of two concepts • the least general generalization of two concepts • the maximally general specialization of two concepts. Define the transitivity of ISA. Define the inheritance of features (including default inheritance and multiple inheritance). G.Tecuci, Learning Agents Laboratory Sample questions Briefly explain the process of reasoning with a plausible version space rule. Define the rule learning problem in Disciple. Briefly describe the rule learning method of Disciple. What is an explanation of an example? Briefly describe analogical reasoning (in general). Briefly describe analogical reasoning in Disciple. Define the rule refinement problem in Disciple. Briefly describe the rule refinement method of Disciple. What is a negative exception? What is a positive exception? Draw a picture representing a plausible version space, as well as a positive example, a negative example, a positive exception and a negative exception. Then briefly define each of these elements. Describe briefly the general architecture of the Disciple shell and the methodology for building a Disciple agent. G.Tecuci, Learning Agents Laboratory Exercise Consider the cells consisting of two bodies, each body having two attributes: - color (that may be yellow or green) and - number of nuclei (1 or 2). The relative position of the bodies is not relevant because they can move inside the cell. a) Indicate ALL the possible generalizations of the following cell, and the generalization relations between them. + ((1 green) (2 yellow)) b) Determine the number of the distinct sets of instances and the number of concept descriptions for this problem. G.Tecuci, Learning Agents Laboratory c) Given the following cell descriptions ((1 green) (1 green)) ((1 yellow) (2 green)) ((1 green) (2 green)) Determine the following minimal generalizations: g(E1, E2), g(E2, E3), g(E3, E1), g(E1, E2, E3) G.Tecuci, Learning Agents Laboratory Exercise The following exercises use the background knowledge consisting of this object hierarchy (semantic network) and the feature definitions from the next slide. SOMETHING ISA INFLAMMABLE -OBJECT ... ISA ISA ISA ADHESIVE ... MATERIAL FRAGIL -OBJECT TOXIC-SUBSTANCE ISA ISA ISA ISA PAPER GLUES GLUES STATE MOWICOLL fluid CONTACT -ADHESIVE ISA PROVIDER GLUES GLUES GLUES ... INSTANCE -OF PROVIDER GLUES MOWICOLL 1 METAL ISA MADE-OF MEMBRANE -OF COLOR INSTANCE MEMBRANE 1 black PART-OF ISA MECHANICAL -CHASSIS MADE-OF INSTANCE -OF MECHANICAL -CHASSIS 1 MADE-OF INSTANCE -OF G.Tecuci, Learning Agents Laboratory ISA CHASSIS MEMBRANE ASSEMBLY ISA INSTANCE -OF CONTAINS ISA CHASSIS MEMBRANE ASSEMBLY 1 MADE-OF ISA GLUE-INC PART-OF LOUDSPEAKER -COMPONENT ISA ISA ISA LOUDSPEAKER ISA ISA ISA CAOUTCHOUC CONTACT -ADHESIVE 1 CHASSIS -ASSEMBLY PART-OF INSTANCE -OF CHASSIS -ASSEMBLY 1 BOLT INSTANCE -OF BOLT1 Feature Definitions name description domain range IS is SOMETHING SOMETHING OBJECT object TASK SOMETHING TO to TASK SOMETHING MADE-OF made of SOMETHING MATERIAL GLUES glues ADHESIVE MATERIAL STATE state SOMETHING {solid fluid gas} TASK task OPERATION TASK INTO into OPERATION TASK ON on TASK SOMETHING PART-OF part of SOMETHING SOMETHING G.Tecuci, Learning Agents Laboratory Exercise Consider the question: “Is there a part of a loudspeaker that is made of metal?” a) Which are all the answers to this question? b) Which are the reasoning operations that need to be performed in order to answer this question. c) Consider one of the answers that requires all these operations and show how the answer is found. G.Tecuci, Learning Agents Laboratory Exercise Consider the following expressions: E1: ?X IS MEMBRANE MADE-OF ?M ?M IS PAPER ?Z IS CONTACT-ADHESIVE GLUES ?M STATE fluid E2: ?X IS MECHANICAL-CHASSIS MADE-OF ?M ?M IS METAL ?Z IS MOWICOLL GLUES ?M a) Find the minimally general generalizations of E1 and E2. b) Find two generalizations of E1 and E2 that are not minimally general generalizations. c) Consider one of the generalizations found at b) and demonstrate why it is a generalization of E1 and E2 but it is not a minimally general generalization. d) What would be a least general generalization of E1 and E2? Does it exist? e) Indicate a specialization of E1. G.Tecuci, Learning Agents Laboratory Exercise Consider the following example and its explanation: IF the task to accomplish is ATTACH OBJECT MEMBRANE1 TO CHASSIS-ASSEMBLY1 THEN accomplish the tasks APPLY OBJECT CONTACT-ADHESIVE1 ON CHASSIS-ASSEMBLY1 PRESS OBJECT MEMBRANE1 ON CHASSIS-ASSEMBLY1 Because CONTACT-ADHESIVE1 IS fluid CONTACT-ADHESIVE1 GLUES PAPER and MEMBRANE1 MADE-OF PAPER CONTACT-ADHESIVE1 GLUES METAL and CHASSIS-ASSEMBLY1 MADE-OF METAL Construct the plausible version space rule learned from them. G.Tecuci, Learning Agents Laboratory Exercise Compose an example analogous with the following one: PAPER MADE-OF GLUES MEMBRANE 1 STATE fluid CONTACT -ADHESIVE 1 GLUES CHASSIS -ASSEMBLY 1 MADE-OF METAL explains IF the task is ATTACH OBJECT MEMBRANE 1TO CHASSIS -ASSEMBLY 1 THEN decompose this task into the subtasks APPLY OBJECT CONTACT -ADHESIVE 1 ON MEMBRANE 1 PRESS OBJECT MEMBRANE 1 ON CHASSIS -ASSEMBLY 1 G.Tecuci, Learning Agents Laboratory Exercise Rule IF the task to accomplish is ATTACH OBJECT ?X TO ?Y Plausible Upper Bound IF ?X IS MADE-OF ?Y IS MADE-OF ?Z IS GLUES GLUES ?M IS ?N IS SOMETHING ?M SOMETHING ?N ADHESIVE ?M ?N MATERIAL MATERIAL Plausible Lower Bound IF ?X IS MADE-OF ?Y IS MADE-OF ?Z IS GLUES GLUES ?M IS ?N IS MEMBRANE1 ?M CHASSIS-ASSEMBLY1 ?N CONTACT-ADHESIVE1 ?M ?N PAPER METAL THEN accomplish the tasks APPLY OBJECT ?Z ON ?X PRESS OBJECT ?X ON ?Y G.Tecuci, Learning Agents Laboratory Find a minimal generalization of the rule that covers the positive example. Positive Example IF the task to accomplish is ATTACH OBJECT BOLT1 TO MECHANICAL-CHASSIS1 THEN accomplish the tasks APPLY OBJECT MOWICOLL1 ON MECHANICAL-CHASSIS1 PRESS OBJECT BOLT1 ON MECHANICAL-CHASSIS1 Exercise IF the task to accomplish is ATTACH OBJECT ?X TO ?Y Rule Plausible Upper Bound IF ?X IS MADE-OF ?Y IS MADE-OF ?Z IS GLUES GLUES ?M IS ?N IS SOMETHING ?M SOMETHING ?N ADHESIVE ?M ?N MATERIAL MATERIAL Plausible Lower Bound IF ?X IS MADE-OF ?Y IS MADE-OF ?Z IS GLUES GLUES ?M IS ?N IS MEMBRANE1 ?M LOUDSPEAKER-COMPONENT ?N LOUDSPEAKER-COMPONENT ?M ?N MATERIAL METAL THEN accomplish the tasks APPLY OBJECT ?Z ON ?X PRESS OBJECT ?X ON ?Y with the positive examples (?X IS MEMBRANE1, ?Y IS CHASSIS-ASSEMBLY1, ?Z IS CONTACT-ADHESIVE1, ?M IS PAPER, ?N IS METAL) (?X IS BOLT1, ?Y IS MECHANICAL-CHASSIS1, ?Z IS MOWICOLL1, ?M IS METAL, ?N IS METAL) G.Tecuci, Learning Agents Laboratory Find a minimal specialization of the rule that does not cover the positive example: • By using an additional explanation of the positive examples; • By empirically specializing the rule. Negative Example IF the task to accomplish is ATTACH OBJECT SCREENING-CAP1 TO LOUDSPEAKER1 THEN accomplish the tasks APPLY OBJECT SCOTCH-TAPE1 ON SCREENING-CAP1 PRESS OBJECT SCREENING-CAP1 ON LOUDSPEAKER1 Exercise Explain how the following questions are answered, and provide the corresponding answer(s): What is the color of membrane? What does contact-adhesive1 glue? Is there a loudspeaker component made of metal? G.Tecuci, Learning Agents Laboratory Exercises The following exercises, marked S1 to S7, are based on the following semantic network from the loudspeaker manufacturing domain: SOMETHING AIR-MOVER CLEANER LOUDSPEAKER -COMPONENT SOFT-CLEANER HARD-CLEANER DAMAGES EMERY-PAPER REMOVES WASTE-MATERIAL AIR-JET-DEVICE SOLVENT ENTREFER MEMBRANE REMOVES REMOVES MAY-HAVE MAY-HAVE AIR-PRESSAIR-SUCKERACETONE ALCOHOL DUST SURPLUS -ADHESIVESURPLUS -PAINT Remark: Consider that each most specific concept, such as DUST or AIR-PRESS, has an instance, such as DUST1 or AIR-PRESS1. G.Tecuci, Learning Agents Laboratory Exercise S1. Consider the following two expressions: E1: ?X ?Y ?Z E2: ?X ?Y ?Z IS REMOVES IS MADE-OF IS SOFT-CLEANER ?Z MEMBRANE ?T WASTE-MATERIAL IS REMOVES NOT-DAMAGES IS MADE-OF IS AIR-SUCKER ?Z PAPER MEMBRANE PAPER DUST Use the generalization rules to show that E1 is more general than E2. G.Tecuci, Learning Agents Laboratory Exercise S2. Determine the generalization of the following two expressions: E1: ?x ?y IS ?z E2: ?x ?y ?z G.Tecuci, Learning Agents Laboratory IS MAY-HAVE IS REMOVES entrefer ?y dust air-sucker ?y IS MAY-HAVE IS IS TYPE REMOVES membrane ?y surplus-adhesive alcohol fluid ?y Exercise S3. Consider the following description: ?z IS REMOVES cleaner surplus-paint Determine all the possible values of ?z. G.Tecuci, Learning Agents Laboratory Exercise S4. Consider the following action description: CLEAN OBJECT OF WITH ?x ?y ?z Condition ?x ?y ?z IS MAY-HAVE IS IS REMOVES entrefer ?y something cleaner ?y Find all the possible values for the variables ?x, ?y and ?z. Indicate some of the corresponding actions. G.Tecuci, Learning Agents Laboratory S5. Consider the following rule: IF the task to perform is CLEAN OBJECT ?x OF ?y Condition ?x IS MAY-HAVE ?y IS ?z IS REMOVES something ?y something cleaner ?y THEN perform the task CLEAN OBJECT ?x OF ?y WITH ?z Describe how this rule is applied to solve the problem: CLEAN OBJECT entrefer1 OF dust1 Which will be the result? Remark: Consider that each most specific concept o from the object ontology has an instance o1. G.Tecuci, Learning Agents Laboratory Exercise S6. Consider the following rule: IF the task to perform is CLEAN OBJECT ?x OF ?y Condition ?x IS MAY-HAVE ?y IS ?z IS REMOVES something ?y something cleaner ?y THEN perform the task CLEAN OBJECT ?x OF ?y WITH ?z Describe how this rule is applied to solve the problem: CLEAN OBJECT membrane1 OF surplus-adhesive1 Which will be the result? Remark: Consider that each most specific concept o from the object ontology has an instance o1. G.Tecuci, Learning Agents Laboratory Exercise S7. Consider the following partially learned rule: Describe how Disciple generalizes this rule so as to cover the following positive example: IF the task to perform is CLEAN OBJECT ?x OF ?y G: plausible upper bound ?x IS something MAY-HAVE ?y ?y IS something ?z IS something REMOVES ?y S: plausible lower bound ?x IS entrefer MAY-HAVE ?y ?y IS dust ?z IS air-sucker REMOVES ?y THEN perform the task CLEAN OBJECT ?x OF ?y WITH ?z IF the task to perform is CLEAN OBJECT membrane1 OF surplus-adhesive1 THEN perform the task CLEAN OBJECT membrane OF surplus-adhesive1 WITH alcohol1 G.Tecuci, Learning Agents Laboratory Exercise Develop an object ontology that represents the following information: Puss is a calico. Herb is a tuna. Charlie is a tuna. All tunas are fishes. All calicos are cats. Cats like to eat fishes. You should define object concepts, object features and instances. G.Tecuci, Learning Agents Laboratory Exercise Develop an object ontology that represents the following information: The color of Apple1 is red. The color of Apple2 is green. Apple1 is an apple. Apple2 is an apple. Apples are fruits. You should define object concepts, object features and instances. G.Tecuci, Learning Agents Laboratory Exercise Develop an object ontology that represents the following information: Basketball players are tall. Muresan is a basketball player. Muresan is tall. You should define object concepts, object features and instances. G.Tecuci, Learning Agents Laboratory Exercise Insert the additional knowledge that platypus lays eggs into the following object ontology: birth-mode mammal subclass-of subclass-of cow Explain the result. G.Tecuci, Learning Agents Laboratory platypus live Exercise Develop an object ontology that represents the following information: "Blue task force 1 penetrates Red mechanized brigade 1 with a force ratio of 10.6. The recommended force ratio for a penetration is 3. A penetration is a complex military task, a military maneuver and a military attack. Use of a penetration indicates that the mission is offensive“ You should draw the ontology and should also define the features used in it (in terms of their domains and ranges). G.Tecuci, Learning Agents Laboratory Exercise Develop an object ontology that represents the following information: "BLUE-TASK-FORCE1 is a blue armored and mechanized infantry battalion assigned to be main effort1. It performs two tasks, penetrate1 and clear1. It has a regular strength and has the following units under its operational control: BLUE-MECHCOMPANY1, BLUE-MECH-COMPANY2, BLUEARMOR-COMPANY1, BLUE-ARMOR-COMPANY2” You should draw the ontology and should also define the features used in it (in terms of their domains and ranges). G.Tecuci, Learning Agents Laboratory Exercise Consider the background knowledge represented by the following generalization hierarchies: any-shape any-color warm-color red orange polygone cold-color yellow blue green triangle black rectangle round circle ellipse square Consider also the following concept: E: ?u IS object COLOR yellow SHAPE circle RADIUS 5 Indicate five different generalization rules. For each such rule determine an expression Eg which is more general than E according to that rule. G.Tecuci, Learning Agents Laboratory Exercise Formalize the following tasks: I need to Identify and test a strategic COG candidate for Okinawa_1945 which is a major theater of war scenario Which is an opposing force in the Okinawa_1945 scenario? US_1945 Therefore I need to Identify and test a strategic COG candidate for US_1945 Is US_1945 a single-member force or a multi-member force? US_1945 is a single-member force Therefore I need to Identify and test a strategic COG candidate for US_1945 which is a single-member force G.Tecuci, Learning Agents Laboratory Exercise Consider the following problem solving episode and its explanation, in the context of the background knowledge the following four slides: War_materiel_and_transports_of_US_1943 is_a_major_generator_of Industrial_capacity_of_US_1943 US_1943 has_as_industrial_factor explains IF the task to accomplish is Identify the strategic COG candidates with respect to the industrial civilization of a force The force is US_1943 THEN A strategic COG relevant factor is strategic COG candidate for a force The force is US_1943 The strategic COG relevant factor is Industrial_capacity_of_US_1943 G.Tecuci, Learning Agents Laboratory a) Find the analogy-based generalization of the explanations and the example. b) Find the plausible version space rule that will be learned from this example. Feature definitions has_as_industrial_factor D: Force R: Industrial_factor is_a_major_generator_of D: Economic_factor R: Product The force is D: task R: Force The strategic COG relevant factor is D: task R: Force G.Tecuci, Learning Agents Laboratory Economic factors Economic_factor Other_ economic_ factor Commerce_ authority Raw_material Strategic_ Raw_ Material Information_ Network_or_system Transportation_ Factor Industrial_ factor Industrial_ authority Industrial_ Center Industrial_ Capacity is_critical_to_ Oil_chromium_ copper_and_bauxite_ the_production_of of_Germany_1943 War_materiel_of has_as_strategic_ _Germany_1943 raw_material is_obtained_from Farm_implement_industry Germany_1943 G.Tecuci, Learning Agents Laboratory Balkans Transportation_ Center Transportation_ Network_or_system industrial_capacity_ of_US_1943 is_a_major_generator_of has_as_industrial_factor war_materiel_and_ transports_of_ US_1943_ US_1943 Farm_implement_industry_of_Italy_1943 Generalization hierarchy of forces <object> Force Group Opposing_force Multi_state_force Single_state_force component_state Anglo_allies_1943 component_state component_state European_axis_1943 component_state Italy_1943 G.Tecuci, Learning Agents Laboratory Multi_group_force US_1943 Britain_1943 Germany_1943 Single_group_force Fragment of the generalization hierarchy <object> Resource_or_ infrastructure_element Resource Product Strategically_essential_resource_ or_infrastructure_element Raw_material Non-strategically_essential goods_or_services Strategically_essential_ goods_or_materiel Strategic_raw_material Farm-implements Strategically_essential_ infrastructure_element War_materiel_and_transports War_materiel_and_fuel Main_airport Sole_airport G.Tecuci, Learning Agents Laboratory Main_seaport Sole_seaport War_materiel_and_fuel_ of_Germany_1943 Farm-implements of_Italy_1943 War_materiel_and_ transports_of_US_1943 Exercise IF Identify the strategic COG candidates with respect to the industrial civilization of a force The force is ?O1 explanation ?O1 has_as_industrial_factor ?O2 ?O2 is_a_major_generator_of ?O3 Plausible Upper Bound Condition ?O1 IS Force has_as_industrial_factor ?O2 ?O2 ?O3 IS Industrial_factor is_a_major_generator_of IS ?O3 Product Minimally generalize the rule to cover the following positive example (considering the background knowledge from the previous four slides): Positive example that satisfies the upper bound IF the task to accomplish is Identify the strategic COG candidates with respect to the industrial civilization of a force The force is Germany_1943 THEN accomplish the task Plausible Lower Bound Condition ?O1 IS US_1943 has_as_industrial_factor ?O2 ?O2 IS Industrial_capacity_of_US_1943 is_a_major_generator_of ?O3 ?O3 IS A strategic COG relevant factor is strategic COG candidate for a force The force is Germany_1943 The strategic COG relevant factor is Industrial_capacity_of_Germany_1943 War_materiel_and_transports_of_US_1943 THEN A strategic COG relevant factor is strategic COG candidate for a force The force is ?O1 The strategic COG relevant factor is ?O2 G.Tecuci, Learning Agents Laboratory explanation Germany_1943 has_as_industrial_factor Industrial_capacity_of_Germany_1943 Industrial_capacity_of_Germany_1943 is_a_major_generator_of War_materiel_and_fuel_of_Germany_1943 Exercise IF Identify the strategic COG candidates with respect to the industrial civilization of a force The force is ?O1 explanation ?O1 has_as_industrial_factor ?O2 ?O2 is_a_major_generator_of ?O3 Plausible Upper Bound Condition ?O1 IS Force has_as_industrial_factor ?O2 ?O2 ?O3 IS Industrial_factor is_a_major_generator_of IS ?O3 Product Plausible Upper Bound Condition ?O1 IS Single_state_force has_as_industrial_factor ?O2 ?O2 ?O3 IS Industrial_capacity is_a_major_generator_of IS ?O3 Strategically_essential_goods_or_materials THEN A strategic COG relevant factor is strategic COG candidate for a force The force is ?O1 The strategic COG relevant factor is ?O2 G.Tecuci, Learning Agents Laboratory Minimally specialize the rule to no longer cover the following negative example (considering the background knowledge from the previous slides): Negative example that satisfies the upper bound IF the task to accomplish is Identify the strategic COG candidates with respect to the industrial civilization of a force The force is Italy_1943 THEN accomplish the task A strategic COG relevant factor is strategic COG candidate for a force The force is Italy_1943 The strategic COG relevant factor is Farm_implement_industry_of_Italy_1943 explanation Italy_1943 has_as_industrial_factor Farm_implement_industry_of_Italy_1943 Farm_implement_industry_of_Italy_1943 is_a_major_generator_of Farm_implements_of_Italy_1943 Repertory grid exercises Define a repertory grid for choosing a course to enroll in. Define a repertory grid for choosing a car. Define a repertory grid for choosing a dissertation director. G.Tecuci, Learning Agents Laboratory Exercise Consider the following two concepts: C 1: ?X IS HEAD COST SCREW HEXAGONAL 5 Indicate different generalization of them. G.Tecuci, Learning Agents Laboratory C 2: ?X IS COST NUT 6 Exercise Consider the following two concepts and ontology. Indicate four specializations of G1 and G2 (including a maximally general specialization). G1: ?X ?M ?Z IS LOUDSPEAKER -COMPONENT G2: ?X MADE-OFM ? ?M IS MATERIAL ?Z IS ADHESIVE GLUES M? IS LOUDSPEAKER -COMPONENT MADE-OFM ? IS MATERIAL IS INFLAMMABLE -OBJECT GLUES M? LOUDSPEAKER -COMPONENT IS MEMBRANE IS IS CHASSIS -ASSEMBLY BOLT ADHESIVE TOXIC-SUBSTANCE INFLAMMABLE -OBJECT IS SCOTCH -TAPE IS IS IS IS IS IS IS -ADHESIVE SUPER-GLUEMOWICOLL CONTACT MATERIAL IS CAOUTCHOUC G.Tecuci, Learning Agents Laboratory IS PAPER IS METAL Exercise Develop an object ontology that represents the following information: Birds have feathers, fly and lay eggs. Albatros is a bird. Donald is a bird. Tracy is an albatros. You should define object concepts, object features and instances. G.Tecuci, Learning Agents Laboratory END G.Tecuci, Learning Agents Laboratory Cooperative problem solving and learning Problem solving with PVS rules Integrated problem solving and learning Demonstration G.Tecuci, Learning Agents Laboratory Generalization by analogy INTELLIGENCE-COLLECTION-MILTARY-TASK RED-CSOP1 INSTANCE-OF TASK SCREEN1 SOVEREIGN-ALLEGIANCE-OF-ORG RED--SIDE explain generalization IF the task to accomplish is: Assess security wrt countering enemy reconnaissance for-coa COA411 THEN accomplish the task: Assess security when enemy recon is present for-coa COA411 for-unit RED-CSOP1 for-recon-action SCREEN1 Knowledge-base constraints on the generalization: Any value of ?O1 should be an instance of: RANGE(FOR-COA) = COA-SPECIFICATION-MICROTHEORY Any value of ?O2 should be an instance of: DOMAIN(TASK) DOMAIN(SOVEREIGN-ALLENGINCE-OF_ORG) RANGE(FOR-UNIT) = MODERN-MILITARY-UNIT--DEPLOYABLE Any value of ?O3 should be an instance of: RANGE(TASK) INTELLIGENCE-COLLECTION-MILITARY-TASK = INTELLIGENCE-COLLECTION-MILITARY-TASK Any value of ?O4 should be an instance of: RANGE(SOVEREIGN-ALLENGINCE-OF_ORG) = ALLEGIANCE-OF-UNIT G.Tecuci, Learning Agents Laboratory A positive example covered by the upper bound Rule: R2 IF the task to accomplish is: Assess-security-wrt-countering-enemy-reconnaissance for-coa ?O1 Question: Is an enemy reconnaissance unit present? Answer: Yes, the enemy unit ?O2 is performing the action ?O3 which is a reconnaissance action. Explanation: Positive example that satisfies the upper bound IF the task to accomplish is: Assess-security-wrt-countering-enemy-reconnaissance for-coa COA421 THEN accomplish the task: Assess-security-when-enemy-recon-is-present for-coa COA421 for-unit RED-CSOP2 for-recon-action SCREEN2 Main Condition Plausible Upper Bound ?O1 IS COA-SPECIFICATION-MICROTHEORY ?O2 IS MODERN-MILITARY-UNIT--DEPLOYABLE SOVEREIGN-ALLEGIANCE-OF-ORG ?O4 TASK ?O3 ?O3 IS INTELLIGENCE-COLLECTION--MILITARY-TASK ?O4 IS ALLEGIANCE-OF-UNIT Plausible Lower Bound ?O1 IS COA411 ?O2 IS MECHANIZED-INFANTRY-UNIT--MIL-SPECIALTY SOVEREIGN-ALLEGIANCE-OF-ORG ?O4 TASK ?O3 ?O3 IS SCREEN1 ?O4 IS RED--SIDE THEN accomplish the task: Assess-security-when-enemy-recon-is-present for-coa ?O1 for-unit ?O2 for-recon-action ?O3 G.Tecuci, Learning Agents Laboratory less general than ?O2 SOVEREIGN-ALLEGIANCE-OF-ORG ?O4 IS RED--SIDE ?O2 TASK ?O3 IS INTELLIGENCE-COLLECTION--MIL-TASK Condition satisfied by positive example ?O1 IS COA421 ?O2 IS MECHANIZED-INFANTRY-UNIT--MIL-SPECIALTY SOVEREIGN-ALLEGIANCE-OF-ORG ?O4 TASK ?O3 ?O3 IS SCREEN2 ?O4 IS RED--SIDE A negative example covered by the upper bound Rule: R$ASWCER-001 IF the task to accomplish is: Assess-security-wrt-countering-enemy-reconnaissance for-coa ?O1 Question: Is an enemy reconnaissance unit present? Negative example that satisfies the upper bound IF the task to accomplish is: Assess-security-wrt-countering-enemy-reconnaissance for-coa COA51 THEN accomplish the task: Assess-security-when-enemy-recon-is-present for-coa COA51 for-unit BLUE-BATTALION1 for-recon-action SCREEN-RIGHT Answer: Yes, the enemy unit ?O2 is performing the action ?O3 which is a reconnaissance action. Explanation: Main Condition Plausible Upper Bound ?O1 IS COA-SPECIFICATION-MICROTHEORY ?O2 IS MODERN-MILITARY-UNIT--DEPLOYABLE SOVEREIGN-ALLEGIANCE-OF-ORG ?O4 TASK ?O3 ?O3 IS INTELLIGENCE-COLLECTION--MILITARY-TASK ?O4 IS ALLEGIANCE-OF-UNIT Plausible Lower Bound ?O1 IS COA-SPECIFICATION-MICROTHEORY ?O2 IS MECHANIZED-INFANTRY-UNIT--MIL-SPECIALTY SOVEREIGN-ALLEGIANCE-OF-ORG ?O4 TASK ?O3 ?O3 IS SCREEN—MILITARY-TASK ?O4 IS RED--SIDE THEN accomplish the task: Assess-security-when-enemy-recon-is-present for-coa ?O1 for-unit ?O2 for-recon-action ?O3 G.Tecuci, Learning Agents Laboratory less general than •?O2 SOVEREIGN-ALLEGIANCE-OF-ORG ?O4 IS RED--SIDE •?O2 TASK ?O3 IS INTELLIGENCE-COLLECTION--MIL-TASK Condition satisfied by positive example ?O1 IS COA51 ?O2 IS BLUE-BATTALION1 SOVEREIGN-ALLEGIANCE-OF-ORG ?O4 TASK ?O3 ?O3 IS SCREEN-RIGHT ?O4 IS BLUE-SIDE ALLEGIANCE-OF-UNIT SUBCLASS-OF BLUE-SIDE _ specialization RED-SIDE INTELLIGENCE-COLLECTION-MILTARY-TASK SUBCLASS-OF SCREEN-MILITARY-TASK INSTANCE-OF SCREEN1 G.Tecuci, Learning Agents Laboratory INSTANCE-OF SCREEN2 COA-SPECIFICATION-MICROTHEORY INSTANCE-OF COA411 INSTANCE-OF COA421