1. Copyright 2003 by The MITRE Corporation 1 Dr. Leo Obrst MITRE Center for Innovative Computing & Informatics Information Semantics Lobrst@mitre.org February 6, 2014February 6, 2014February 6, 2014 Ontologies for Semantically Interoperable Systems Semantic Representation Semantic Mapping Semantic Interoperability

2. 2 Overview The Problem Tightness of Coupling & Explicit Semantics Semantic Integration Implies Semantic Composition Dimensions of Interoperability & Integration Ontologies –The Ontology Spectrum –What are Ontologies? –Levels of Ontology Representation –What Problems do Ontologies Help Solve? Ontologies for Semantically Interoperable Systems –Enabling Semantic Interoperability –Examples –Visions –What do We Want the Future to be?

3. 3 The problem With the increasing complexity of our systems and our IT needs, and the distance between systems, we need to go toward human level interaction We need to maximize the amount of semantics we can utilize and make it increasingly explicit From data and information level, we need to go toward human semantic level interaction DATAInformationKnowledge Run84 ID=08 NULL PARRT ACC ID=34 e 5 & # ~ Q ü @ ¥ ¥ Æ Å Tank ¥ Noise Human Meaning Vehicle Located at Semi-mountainous terrain obscured decide Vise maneuver Semantic representation & semantic interoperability/integration become very important

4. 4 Tightness of Coupling & Semantic Explicitness Data Application Implicit, TIGHT Explicit, Loose 1 System: Small Set of Developers Local Far Same Process Space Same Address Space Same CPU Same OS Same Programming Language Same DBMS Same Local Area Network Systems of Systems Enterprise Community Internet Same Wide Area Network Same Intranet Federated DBs Data Warehouses Data Marts Ontologies Linking Libraries OOP Agent Programming Web Services: SOAP Distributed Systems Applets Semantic Mappings Semantic Brokers Looseness of Coupling Semantics Explicitness From Local, Tight, Implicit To Far, Loose, Explicit XML Conceptual Models RDF/S, OWL Web Services: UDDI, WSDL OWL-S Modal Policies

5. 5 Semantic Interoperability: Tight to Loose Coupling Tight coupling: applies to databases, systems –Same address space, same process space, same operating system, same machine –Semantic compacts can be made because semantics stays in the minds of the developers who agree Loose coupling –Different platforms, networks, anywhere on Internet –Semantics must be explicit: agents, programs need to interpret the semantics directly, to interoperate semantically –Levels: systems of systems, enterprise, community, value chains/pipes Ontologies (explicitly represented, logical semantics): increasingly needed the higher you go

6. 6 Semantic Integration Implies Semantic Composition Simple Procedure Integration & Composition Concatenation, alignment of calling Procedure with called procedure: Caller: Do_this (integer: 5, string: sales) Called: Do_this (integer: X, string: Y) Simple Syntactic Object Integration & Composition Alignment of embedded interface definition language statements mapping two CORBA, Javabean objects Simple Semantic Model, Knowledge Integration & Composition Unification of tree or graph structures, with reasoning, simple Semantic Web ontologies: - signifies the composition operation Complex Semantic Model, Knowledge, System Integration & Composition Unification of complex networks of graph Structures, with complex reasoning, complex Semantic Web ontologies: 1960 1998 2005 2010

7. 7 Dimensions of Interoperability & Integration Enterprise Object Data System Application Component 0%100% 6 Levels of Interoperability 3 Kinds of Integration Interoperability Scale Our interest lies here Community

8. 8 Semantic Interoperability/Integration Definition To interoperate is to participate in a common purpose –Operation sets the context –Purpose is the intention, the end to which activity is directed Semantics is fundamentally interpretation –Within a particular context –From a particular point of view Semantic Interoperability/Integration is fundamentally driven by communication of purpose –Participants determined by interpreting capacity to meet operational objectives –Service obligations and responsibilities explicitly contracted

9. 9 weak semantics strong semantics Is Disjoint Subclass of with transitivity property Modal Logic Logical Theory Thesaurus Has Narrower Meaning Than Taxonomy Is Sub-Classification of Conceptual Model Is Subclass of DB Schemas, XML Schema UML First Order Logic Relational Model, XML ER Extended ER Description Logic DAML+OIL, OWL RDF/S XTM Ontology Spectrum: One View Syntactic Interoperability Structural Interoperability Semantic Interoperability

10. 10 Logical Theory Thesaurus Has Narrower Meaning Than Taxonomy Is Sub-Classification of Conceptual Model Is Subclass of Is Disjoint Subclass of with transitivity property weak semantics strong semantics DB Schemas, XML Schema UML Modal Logic First Order Logic Relational Model, XML ER Extended ER Description Logic DAML+OIL, OWL RDF/S XTM Ontology Spectrum: One View Problem: Very General Semantic Expressivity: Very High Problem: Local Semantic Expressivity: Low Problem: General Semantic Expressivity: Medium Problem: Local Semantic Expressivity: High Syntactic Interoperability Structural Interoperability Semantic Interoperability

11. 11.251.25SquareXAB035.75 1.5 RoundXAB023 …Price ($US) Size (in) ShapeCatalo g No..4531S 55029 8.3537R 55029 6 … Price ($US) Diam (mm) Geom. Part No. Washer Catalog No. Shape Size Price iMetal Corp. E-Machina iMetal Corp. E-Machina Manufactur er.451.25Square550298.351.5Round550296.751.5RoundXAB023.251.25SquareXAB035 … Price ($US) Size (in) ShapeMfr No. Supplier A Supplier B Buye r Ontology A Business Example of Ontology

12. 12 Architecture: Ontology & Applications Ontology Layer Ontology Application Services Layer Application Layer User Interface Layer Semantic Representation Requirements User (& presentation) Requirements Support for User to Representation Requirements SearchTransact User Roles Buyer (Engr., Analyst) Seller SearchTransactNavigate Get/Put Data Make/Get AliasLook-upContextualizeInfer Taxonomies MetadataAttributes In the emerging Web Services paradigm, Levels 1 - 4 consist of composable services 1 2 3 4

13. 13 What Problems Do Ontologies Help Solve? Heterogeneous database problem –Different organizational units, Service Needers/Providers have radically different databases –Different syntactically: whats the format? –Different structurally: how are they structured? –Different semantically: what do they mean? –They all speak different languages (access, description, schemas, meaning) –Integration: rather than N 2 problem, with single, adequate Ontology reduces to N Enterprise-wide system interoperability problem Enterprise-wide system interoperability problem –Currently: system-of-systems, vertical stovepipes –Ontologies act as conceptual model representing enterprise consensus semantics Relevant document retrieval/question-answering problem –What is the meaning of your query? –What is the meaning of documents that would satisfy your query? –Can you obtain only meaningful, relevant documents?

14. 14 Enabling Semantic Interoperability Semantic Interoperability is enabled through: –Establishing base semantic representation via ontologies (class level) and their knowledge bases (instance level) –Defining semantic mappings & transformations among ontologies (and treating these mappings as individual theories just like ontologies) –Defining algorithms that can determine semantic similarity and employing their output in a semantic mapping facility that uses ontologies The use of ontologies & semantic mapping software can reduce the loss of semantics (meaning) in information exchange among heterogeneous applications, such as: –Web Services –E-Commerce, E-Business –Enterprise architectures, infrastructures, and applications –Complex C4ISR systems-of-systems –Integrated Intelligence analysis

15. 15 Semantic Interoperability, Integration: Multiple Semantics Multiple contexts, views, application & user perspectives Multiple levels of precision, specification, definiteness required Multiple levels of semantic model verisimilitude, fidelity, granularity Multiple kinds of semantic mappings, transformations needed: –Entities, Relations, Properties, Ontologies, Model Modules, Namespaces, Meta-Levels, Facets (i.e., properties of properties), Units of Measure, Conversions, etc.

16. 16 Simple Example: Semantics of Date Across Applications System 1 Instance of Concept: Date 1 –Attribute: YR = Int 1 –Attribute: MO = String Aug –Attribute: DY = Int 12 System 2 : Instance of Concept = Date 2 –Attribute: DayOfWeek = Sunday –Attribute: ActualDate = String 12082001 Semantically Equivalent? Then How? DATE 2 DayOfWeek ActualDate DATE 1 MO YR DY Exactly Semantically Equivalent to? No: Approximately Semantically Equivalent to. So Mappings and Transformations are Needed! Add Assertions, Apply Transformations (directional) Once Assertions, Transformations Defined: become part of Integration Ontology & Reused Date2.ActualDate Date1.DY Date1.MO Date1.YR

17. 17 Simple Example: Semantics of Location Across Applications System 1 Instance of Concept: Location 1 –Attribute: SourceDeadReckoning = A –Attribute: SourceDRLatitude = B –Attribute: SourceDRLongitude = C –Attribute: TargetDRBearingLine = D –Attribute: TargetDRAltitude = E –Attribute: ActualMeasuredAltitude = F –Attribute: PositionLine = G System 2 : Instance of Concept: Location 2 –Attribute: Address = H –Attribute: City = I –Attribute: StateProvince = J –Attribute: Country = K –Attribute: MailCode = L Approximately Semantically Equivalent to?

18. 18 Electronic Commerce Example: One Company Products Metal Health Electronic Chemical Distributor Manufacturer Wholesaler Retailer EndRun TradingPartners TransWorld iMicro 3Initial Location Africa Europe Spain Portugal Asia Time Point Interval Coordinate System UTM Geographic LatLong GPS UnitOfMeasure Distance Mass Liquid Solid Shipping Methods Air Ground Truck RegionalCarrier LocalCarrier Sea Applications TradingHub RFI/RFQ Sell ShippedBy ObtainedFrom LocatedAt GivenBy MeasuredBy Uses Support AvailableAt Train

19. 19 Now Assume Each Company Has Separate Enterprise Semantics, Multiply by the Number of Companies, & Have Them Interoperate and Preserve Semantics Try doing this without Ontologies! You can, but its a Nightmare, and it COSTS: Now & Later!

20. 20 Emerging XML Stack Architecture for the Semantic Web + Grid + Agents Semantic Brokers Intelligent Agents Advanced Applications Use, Intent: Pragmatics Trust: Proof + Security + Identity Reasoning/Proof Methods OWL, DAML+OIL: Ontologies RDF Schema: Ontologies RDF: Instances (assertions) XML Schema: Encodings of Data Elements & Descriptions, Data Types, Local Models XML: Base Documents Grid & Semantic Grid: New System Services, Intelligent QoS Sem-Grid ServicesWater, LISP? Syntax: Data Structure Semantics Higher Semantics Reasoning/Proof XML XML Schema RDF/RDF Schema OWL Inference Engine Trust Security/Identity Use, Intent Pragmatic Web Intelligent Domain Services, Applications Agents, Brokers, Policies

21. 21 Semantic Web Services Stack Adapted from: Bussler, Christoph; Dieter Fensel; Alexander Maedche,. 2003. A Conceptual Architecture for Semantic Web Enabled Web Services. Semantics Pragmatics

22. 22 Z Y XW VT S A M IJ BCDE F GH Simple, Informal E-Commerce Application Taxonomy (Reference) Ontology Well-defined subclass relation Other ontological relations Ill-defined parent-child relation Mappings Industrial process Products of the process Equipment used In the process Employees involved in the process Generated from Industrial process Specific product General product Semantic Mappings

23. 23 Electronics Namespace UNSPSC Namespace Ontology Ontology node UNSPSC node Inheritance (subclass) Taxonomic Standard to Ontology Mapping: e.g., Web Services UNSPSC Mapped to Electronics Domain Ontology Use of Nebenstruktur (shadow structure)

24. 24 Z Y X WVTS A M IJ BCDE F GH Ontology subclass relation Other ontological relations Application subclass relation Mappings (equivalence) Implementing Mappings: Semantic Model of Application in the Ontology Model of Application: All ontology-relevant application structures are included in the ontology model –Create mapping relation from an ontology node to nodes in the application model Claw hammer Hammer as used in carpenters catalog X

25. 25 Semantic Issues: Complexity An ontology allows for near linear semantic integration (actually 2n-1) rather than near n 2 (actually n 2 - n) integration –Each application/database maps to the "lingua franca" of the ontology, rather than to each other AC AB BC ACB Ordinary Integration Ontology Integration AD BD CD Add D: AD AB CD BC A D 2 Nodes 3 Nodes 4 Nodes 5 Nodes 2 Edges 6 Edges 12 Edges 20 Edges 2 Nodes 3 Nodes 4 Nodes 5 Nodes 2 Edges 4 Edges 6 Edges 8 Edges

26. 26 Vision: Semantic Broker Web-Based Machine-Interpretable Semantics (stacked languages) Use/Intent Proof OWL Agent Services Web Services RDF/S XTM XLT Specific XML Languages XML Schema XML Schema Application Data Mappings Ontologies Documents Application Schema Application Data Application Schema Application Data Application Semantic Broker Semantic Mapper Contexts Requests Services

27. 27 Vision: Semantically Interoperable Systems Semantic Broker Active Application Agent Active Application Agent Active Application Agent Application Users: Purchasers, Sellers, Decision-Makers Consumers, Analysts, Manufacturers Application Meta-data Agency Meta-data Meta-Knowledge Upper Ontology: Generic Base Organizations Interaction Knowledge WorkflowProcesses Mapping Knowledge Products & Svcs Ontologies Fielded Systems Semantic Mappings Queries Ontology and Reasoning Services Databases Documents

28. 28 What do we want the future to be? 2100 A.D: models, models, models There are no human-programmed programming languages There are only Models Ontological Models Knowledge Models Belief Models Application Models Presentation Models Target Platform Models Transformations, Compilations Executable Code INFRASTRUCTUREINFRASTRUCTURE

29. 29 Contact Questions? lobrst@mitre.org Shameless Plug: The Semantic Web: The Future of XML, Web Services, and Knowledge Management, -- Mike Daconta, Leo Obrst, & Kevin Smith, Wiley, June, 2003 http://www.amazon.com/exec/obidos/ASIN/0471432571/qid%3D10502646 00/sr%3D11-1/ref%3Dsr%5F11%5F1/103-0725498-4215019 Contents: 1.What is the Semantic Web? 2.The Business Case for the Semantic Web 3.Understanding XML and its Impact on the Enterprise 4.Understanding Web Services 5.Understanding the Resource Description Framework 6.Understanding the Rest of the Alphabet Soup 7.Understanding Taxonomies 8.Understanding Ontologies 9.Crafting Your Companys Roadmap to the Semantic Web

30. 30 Backup

31. 31 Ontology & Ontologies 1 An ontology defines the terms used to describe and represent an area of knowledge (subject matter) –An ontology also is the model (set of concepts) for the meaning of those terms –An ontology thus defines the vocabulary and the meaning of that vocabulary Ontologies are used by people, databases, and applications that need to share domain information –Domain: a specific subject area or area of knowledge, like medicine, tool manufacturing, real estate, automobile repair, financial management, etc. Ontologies include computer-usable definitions of basic concepts in the domain and the relationships among them –They encode domain knowledge (modular) –Knowledge that spans domains (composable) –Make knowledge available (reusable)

32. 32 Ontology & Ontologies 2 The term ontology has been used to describe models with different degrees of structure (Ontology Spectrum) –Less structure: Taxonomies (Semio taxonomies, Yahoo hierarchy, biological taxonomy), Database Schemas (many) and metadata schemes (ICML, ebXML, WSDL) –More Structure: Thesauri (WordNet, CALL, DTIC), Conceptual Models (OO models, UML) –Most Structure: Logical Theories (Ontolingua, TOVE, CYC, Semantic Web) Ontologies are usually expressed in a logic-based language –Enabling detailed, sound, meaningful distinctions to be made among the classes, properties, & relations –More expressive meaning but maintain computability Using ontologies, tomorrow's applications can be "intelligent – Work at the human conceptual level Ontologies are usually developed using special tools that can model rich semantics

33. 33 Ontology & Ontologies 3 Ontologies are developed by a team –Domain Experts: have the domain knowledge –Ontologists: know how to formally model knowledge, semantics On-going research investigates semi-automation of ontology development –State-of-art for quite some time will be semi-automation –Humans have rich semantic models & understanding, machines poor so far –Want our machines to interact more closely at human concept level –The more & richer the knowledge sources developed & used, the easier it gets (bootstrapping, learning) Rigorous ontology development methodologies evolving (e.g., Methontology) Tools emerging to assist domain experts in building ontologies (OntoClean)

34. 34 Axioms, Inference Rules, Theorems, Theory Theory Theorems Licensed by a valid proof using inference rules Possible other theorems (as yet unproven) Axioms

35. 35 AxiomsInference RulesTheorems Class(Thing) Class(Person) Class(Parent) Class(Child) If SubClass(X, Y) then X is a subset of Y. This also means that if A is a member of Class(X), then A is a member of Class(Y) SubClass(Person, Thing) SubClass(Parent, Person) SubClass(Child, Person) ParentOf(Parent, Child) NameOf(Person, String) AgeOf(Person, Integer) If X is a member of Class (Parent) and Y is a member of Class(Child), then (X Y) And-introduction: given P, Q, it is valid to infer P Q. Or-introduction: given P, it is valid to infer P Q. And-elimination: given P Q, it is valid to infer P. Excluded middle: P P (i.e., either something is true or its negation is true) If P Q are true, then so is P Q. If X is a member of Class(Parent), then X is a member of Class(Person). If X is a member of Class(Child), then X is a member of Class(Person). If X is a member of Class(Child), then NameOf(X, Y) and Y is a String. If Person(JohnSmith), then ParentOf(JohnSmith, JohnSmith).

36. 36 Ontology Representation Levels

37. 37 E-commerce Area of Interest Mostly This Middle Ontology (Domain-spanning Knowledge) Most General Thing Upper Ontology (Generic Common Knowledge) Products/Services Processes Organizations Locations Lower Ontology (individual domains) Metal Parts Art Supplies Lowest Ontology (sub-domains) Washers But Also This! Ontology: General Picture at Object Level