Ontologies and Semantic Networks Sven Koenig, USC Russell and - PDF document

12/18/2019 Ontologies and Semantic Networks Sven Koenig, USC Russell and Norvig, 3 rd Edition, Section 12.5.1 These slides are new and can contain mistakes and typos. Please report them to Sven (skoenig@usc.edu). 1 Ontology • Ontology = a model for describing the world that consists of a set of types, properties, and relationship types 2 1

12/18/2019 Example: Taxonomic Knowledge • “All office machines get their energy from wall outlets.” • “All printers are office machines.” • “All laser printers are printers.” • “Hobbes is a laser printer.” 3 Example: Taxonomic Knowledge • Knowledge base in first-order logic • FORALL x IsOfficeMachine(x) IMPLIES EnergySource(x, WallOutlet) • FORALL x IsPrinter(x) IMPLIES IsOfficeMachine(x) • FORALL x IsLaserPrinter(x) IMPLIES IsPrinter(x) • IsLaserPrinter(Hobbes) • We can use resolution to show that the knowledge base entails • EnergySource(Hobbes, WallOutlet) • But the knowledge base and resolution are difficult to understand by non-experts and resolution is often slow (and non-trivial to implement), so we are looking for alternative ways to represent knowledge and reason with it. 4 2

12/18/2019 Semantic Networks EnergySource WallOutlets OfficeMachines isa isa Robots Printers isa isa isa isa DeliveryRobots CleaningRobots LaserPrinters InkjetPrinters isa isa R2D2 Hobbes 5 Semantic Networks EnergySource WallOutlets OfficeMachines subset subset Robots Printers subset subset subset subset DeliveryRobots CleaningRobots LaserPrinters InkjetPrinters element element R2D2 Hobbes 6 3

12/18/2019 Semantic Networks subset subset A B FORALL x (A(x) IMPLIES B(x)) Cats Mammals element element A B B(A) Bill Cats R Age A B R(A,B) Bill 12 Legs R Birds 2 A B FORALL x (A(x) IMPLIES R(x,B)) Parent R A B Birds Birds FORALL x (A(x) IMPLIES EXISTS y (B(y) AND R(x,y))) 7 Semantic Networks • How would you depict “R2D2 is not a cleaning robot”? • How would you depict “R2D2 is a delivery or cleaning robot”? 8 4

12/18/2019 Semantic Networks • A special purpose reasoning procedure (“pointer following”) makes reasoning about properties easy, using the inheritance of properties. 9 Semantic Networks • What’s the energy source of Hobbes? EnergySource WallOutlets OfficeMachines subset subset Robots Printers subset subset subset subset DeliveryRobots CleaningRobots LaserPrinters InkjetPrinters element element R2D2 Hobbes 10 5

12/18/2019 Semantic Networks • “Yesterday, I looked out of the window and saw a bird.” • Do you think that the bird I saw could (likely) fly? • Why do people jump to conclusions here? They reason with defaults. • “Let me continue. It had a broken wing and sat on the ground.” • If you thought that the bird could fly, you now need to revise your conclusion. • This cannot be done (in straight-forward ways) with first-order logic since first-order logic is monotonic, meaning that KB AND KB’ ⊨ S whenever KB ⊨ S. • For example, IsBird(Tweety) AND BrokenWing(Tweety) ⊨ CanFly(Tweety) if IsBird(Tweety) ⊨ CanFly(Tweety). • However, default reasoning can be done easily with semantic networks. 11 Semantic Networks • “R2D2 is an office machine.” EnergySource WallOutlets OfficeMachines subset subset EnergySource Batteries Robots Printers subset subset subset subset DeliveryRobots CleaningRobots LaserPrinters InkjetPrinters element element R2D2 Hobbes 12 6

12/18/2019 Semantic Networks • “Let me continue. In fact, R2D2 is a delivery robot.” EnergySource WallOutlets OfficeMachines subset subset EnergySource Batteries Robots Printers subset subset subset subset DeliveryRobots CleaningRobots LaserPrinters InkjetPrinters element element R2D2 Hobbes 13 Semantic Networks • Semantic networks can represent stories. • “John gave a book to Mary.” Book Give Event element element agent object Book 23 John Event 7 beneficiary Mary 14 7

12/18/2019 Semantic Networks • Spreading activation (“marker passing”) can reason about stories represented with semantic networks. • Activation is the arousal level of a node. • Nodes mentioned in the story are activated. • Whenever a node receives activation, a fraction of that activation spreads with or against the semantic links connected to the node. • The higher the activation of a node, the more easily the corresponding concept can be accessed from memory. 15 This is a good time to point out the Statement on Academic Conduct and Support Systems on the class Warning webpage, which includes resources for suicide prevention! • We use a suicide prevention example on the next couple of slides for historic reasons: early AI researchers studied similar examples. Chapter 6 by Hendler in “Models of Cognition: A Review of Cognitive Science” edited by Sharkey (page 147) 16 8

12/18/2019 Semantic Networks • Background knowledge is represented as a huge semantic network. Rope Knife Tool Tool Stabbing Hanging Bankruptcy Terminally Ill Oneself Event Oneself Event Event Event subset subset Causes Causes Precondition Severe Suicide Event Depression 17 Semantic Networks • “John went bankrupt. He bought a rope.” object agent Rope 1 Event 2 John element agent Knife Rope Event 1 element Tool Tool Hanging Stabbing Bankruptcy Terminally Ill Oneself Event Oneself Event Event Event subset subset Causes Causes Precondition Severe Suicide Event Depression 18 9

12/18/2019 Buying Event Semantic Networks element subset Event • “John went bankrupt. He bought a rope.” subset subset subset object agent Rope 1 Event 2 John element agent Knife Event 1 Rope element Tool Tool Hanging Stabbing Bankruptcy Terminally Ill Oneself Event Oneself Event Event Event subset subset Causes Causes Precondition Severe Suicide Event Depression 19 Frames (very similar to semantic networks) Printers SubsetOf: OfficeMachines SupersetOf: {LaserPrinters, InkjetPrinters} EnergySource: WallOutlet Creator: Sven Koenig Date: Sep 10, 2018 Hobbes ElementOf: LaserPrinters EnergyConsumption: 180 Watt/hour spreadsheet-type calculation EnergyCostPerHour: Utility:EnergyCost * this:EnergyConsumption via “procedural attachment” Creator: Sven Koenig meta information Date: Feb 23, 2015 20 10

12/18/2019 Semantic Networks and Frames • Properties (some versus first-order logic) • Knowledge base (appears) easy to understand by humans but semantics is often not well defined in practice • Problems with multiple inheritance of incompatible properties • More expressive than first-order logic with regard to default reasoning and procedural attachments • Less expressive (or more complicated) than first-order logic with regard to some logical operators such as negation and disjunction • Reasoning easy to implement and efficient but limited in capability due to special-purpose reasoning procedures • Some reasoning (such as inheritance) is easy to explain to non-experts 21 11