Скачать презентацию Knowledge Representations for Semantic Interoperability Conceptual Primitives Knowledge Скачать презентацию Knowledge Representations for Semantic Interoperability Conceptual Primitives Knowledge

e13a4edcec5368d2537d059f6be6a283.ppt

  • Количество слайдов: 27

Knowledge Representations for Semantic Interoperability Conceptual Primitives Knowledge Structures Reasoning Mechanisms Tom Beckman Principal, Knowledge Representations for Semantic Interoperability Conceptual Primitives Knowledge Structures Reasoning Mechanisms Tom Beckman Principal, Beckman Associates [email protected] net 202 -362 -5774 © 2006 Tom Beckman

Outline l l l l l Forms of Knowledge Representation Dimensions Content Knowledge Forms Outline l l l l l Forms of Knowledge Representation Dimensions Content Knowledge Forms Concept Dimensions Conceptual Primitives Knowledge Structures Reasoning Mechanisms Generic Tasks Semantic Web Services © 2006 Tom Beckman

Introduction l Artificial Intelligence Disciplines Applied to Semantics Ø Ø Ø l Explicit Representation Introduction l Artificial Intelligence Disciplines Applied to Semantics Ø Ø Ø l Explicit Representation of: Ø Ø Ø l Expert Systems Linguistics and Natural Language Understanding Machine Learning Knowledge Experience Expertise Knowledge Representation Categories: Ø Ø Semantic: Symbols, numbers, language, meaning Sensory: Images and signals are interpreted into symbolic form © 2006 Tom Beckman

Knowledge Representation Basics Knowledge Representation Characteristics: Ø Ø Models knowledge and reasoning about knowledge Knowledge Representation Basics Knowledge Representation Characteristics: Ø Ø Models knowledge and reasoning about knowledge Describes characteristics and dimensions of knowledge Formally defines structures and processes for electronic and human reasoning Exposes Knowledge Structures and hides Inference Engines Symbolic Knowledge Representation is comprised of two parts: Ø Ø Knowledge Structures: Reasoning Mechanisms: Objects Process Declarative Procedural Static Dynamic Knowledge Structures consist of symbols that explicitly define, describe, organize, and link knowledge Ø Ø Nodes: Symbols as concepts Links: Relations between symbols Reasoning Mechanisms process Knowledge Structures in order to: Ø Ø Solve problems, create knowledge, explain reasoning & results Calculate measures of uncertainty and importance to improve © 2006 Tom Beckman

Forms of Knowledge Representation l Numeric: Ø Ø Ø l Symbolic: Ø Ø l Forms of Knowledge Representation l Numeric: Ø Ø Ø l Symbolic: Ø Ø l Less precise but most concise Ease of reasoning and explanation Linguistic: Ø l Most precise reasoning Needs explanation of computing and results Often hidden – working in the background behind symbols and text Best for explanation and understanding Not as precise, and not concise Hard to directly reason with – language parsers Image: Described and classified using taxonomies and metadata Signal: Time serieds are described, interpreted, and classified using taxonomies, metadata, and analysis engines © 2006 Tom Beckman

Knowledge Representation Dimensions l Concept: Ø Ø Ø l Structure: Ø Ø Ø l Knowledge Representation Dimensions l Concept: Ø Ø Ø l Structure: Ø Ø Ø l Symbolic Format: Concept Types: Object, Entity, and Abstraction Domain content knowledge Declarative representation Composed of Nodes and Links Expert System types parallel human cognitive schema Process: Ø Ø Ø Procedural representation Reasoning and Inference Modeling and Simulation © 2006 Tom Beckman

Concept Dimensions l l l Meaning: is nothing more than the sum of these Concept Dimensions l l l Meaning: is nothing more than the sum of these concept dimensions Definition Attributes: Ø Ø l Relations: Ø Ø l Ø Part of speech Grammar rules Context: Ø Ø l Between concepts Between attributes Linguistics: Ø l Stereotypical description of characteristics Format: Based on user experience and purpose Common understanding Mental Models and Cognitive Schema © 2006 Tom Beckman

Concept Typology l Objects: Ø Ø Ø l Entities: Ø Ø Ø Ø l Concept Typology l Objects: Ø Ø Ø l Entities: Ø Ø Ø Ø l Inanimate physical objects with characteristics: size, shape, color Man-made and naturally occurring objects Behaviors obey natural laws of physics and chemistry Animate systems possess purpose and goals Types: Humans, animals, electronic agents, and plants Sense, reason, and take actions Humans and animals have emotions, values, beliefs, and drives Possess behaviors, procedures, and methods Have resources: memory, knowledge, skills Communicate with other intelligent entities Abstraction: Ø Ø Created by entities to describe, order, and classify the world, perform tasks, and model systems Represent and model real, mental, and virtual worlds Semantic and symbol based representations Obeys laws of inference in closed systems © 2006 Tom Beckman

Conceptual Primitives Knowledge Templates are key conceptual primitives: Ø Ø Represent assertions – the Conceptual Primitives Knowledge Templates are key conceptual primitives: Ø Ø Represent assertions – the basic building blocks of structures Define, describe, and detail symbol features, values, & relations Can also represent Uncertainty & Importance Come in several standard templates: l l l Basic: Faceted: Measured: Declarative Conceptual Primitives: Ø Ø Feature Descriptor: Ex: Relation: Ex: Procedural Conceptual Primitives: Ø Ø Action: Ex: Inference: Ex: These knowledge elements have certain properties: Ø Ø Ø Naming (Object) Describing (Attribute and Value) Organizing (Hierarchy) Relating (Functional, Causal, & Empirical Links) Constraining and Negating (Networks and Rules) © 2006 Tom Beckman

Attribute Value Typology Numeric: Ø Ø Ø Ordinal: Likert Scale Interval: Range, Continuous Variable Attribute Value Typology Numeric: Ø Ø Ø Ordinal: Likert Scale Interval: Range, Continuous Variable Continuous: Ratio, normalized continuous variable Semantic: Ø Text Value Types: Ø Ø Unstructured: Instant Messaging Semi-Structured: Email, Memo Structured: Document, Hypertext Symbolic Value Types: l l l Binary: Boolean Categorical: Unrelated Nominal Ordinal: Related Nominal Sensory: Ø Ø Image: Digital spatial array, picture, video Signal: Time series, audio, sensor © 2006 Tom Beckman

Types of Concept Relations l l l l l Synonym and Antonym Typing and Types of Concept Relations l l l l l Synonym and Antonym Typing and Metadata Hierarchy: Taxonomy and Ontology Composition: Object parts Network: BBN, NN, Fuzzy Sets, Semantic, & Constraint Causal: Inference chains Association: Statistical and Bayesian Temporal: Model/Process of ordered activities Physical: Location, relative physical arrangement © 2006 Tom Beckman

Expert System Typology l l l l Case or Similarity Rule Object Network Process/Model Expert System Typology l l l l Case or Similarity Rule Object Network Process/Model Hybrid Expert Systems are explicit representations of human cognitive schema © 2006 Tom Beckman

Reasoning Mechanism Typology l l l l Document Search: Keyword Bayesian search Database Query: Reasoning Mechanism Typology l l l l Document Search: Keyword Bayesian search Database Query: Relational and Dynamic Queries Web Query: Keyword Search, Semantic Search Engines: Brute Force, Beam, Best-First Similarity-Based Reasoning: Cases Forward & Backward Chaining: Rules Graph Reasoning: Networks and Decision Tree Logic: Propositional, FOPC, Fuzzy Statistical and Bayesian Reasoning Object Methods: Inheritance and Classification Simulation and Modeling: Process Concept Classification: Metadata and Metatagging Natural Language Understanding Analysis Methods: Data Mining, Text Mining, & Knowledge Discovery Machine Learning © 2006 Tom Beckman

Semantic Web Components l l Domain Content: Knowledge, experience, and expertise Domain Taxonomy and Semantic Web Components l l Domain Content: Knowledge, experience, and expertise Domain Taxonomy and Ontology: Ø Ø l l Organization and Structure: Web sites and document collections Classification Methods: Ø Ø l l l RDF/OWL Object Methods: Inheritance and Classification Similarity-based Rule-based Network-based Object-based Indexing: Item typing and meta-tagging Linguistics: Natural Language Processing & Text Generation Search Query: Keyword Bayesian Search & Semantic Search Entity Extraction: People, places, and events Analysis Methods: Data Mining, Text Mining, Link Analysis, Machine Learning, & Knowledge Discovery Intelligent Agents: Simulation and Modeling © 2006 Tom Beckman

Methods to Increase the Value of Knowledge Apply Knowledge Create Knowledge Capture Knowledge Organize Methods to Increase the Value of Knowledge Apply Knowledge Create Knowledge Capture Knowledge Organize Knowledge Share Knowledge Absorb Knowledge Improve Quality of Knowledge © 2006 Tom Beckman

Apply Knowledge Perform a Task Manage a Task Make a Decision Solve a Problem Apply Knowledge Perform a Task Manage a Task Make a Decision Solve a Problem Improve Task Performance © 2006 Tom Beckman

Create Knowledge Create a new KM framework: l l l Structure/Taxonomy Content/Knowledge Function/Process/Method System Create Knowledge Create a new KM framework: l l l Structure/Taxonomy Content/Knowledge Function/Process/Method System Outputs/Performance Resources Organize and combine existing knowledge Synthesize new knowledge through research and analysis Innovate towards a stretch goal Brainstorm and other directed creativity © 2006 Tom Beckman

Capture Knowledge Identify knowledge sources: l Organizational Documents l Books, Journals, and Internet l Capture Knowledge Identify knowledge sources: l Organizational Documents l Books, Journals, and Internet l Internal Subject Matter Experts l External Consultants and Experts Define key knowledge subjects Elicit key knowledge from internal SMEs Develop core competencies © 2006 Tom Beckman

Share Knowledge Communication Media: l l Web Sites and Email Publications Meetings Communities Define Share Knowledge Communication Media: l l Web Sites and Email Publications Meetings Communities Define audiences and context Organize communities of practice and interest Hold workshops and seminars to solve problems, make decisions, and develop new knowledge © 2006 Tom Beckman

Absorb Knowledge Learn new concepts and practices: l l Read books and journals Take Absorb Knowledge Learn new concepts and practices: l l Read books and journals Take classroom training courses Gain experience Peer and community seminars Learn as groups and individuals Identify best practices Organize and combine existing knowledge © 2006 Tom Beckman

Improve the Quality of Knowledge Transform Knowledge Up the Value Taxonomy Capability Expertise Knowledge Improve the Quality of Knowledge Transform Knowledge Up the Value Taxonomy Capability Expertise Knowledge Information Data Sensory © 2006 Tom Beckman

Valuation of Knowledge Assets Knowledge is intangible and difficult to measure/value Most knowledge loses Valuation of Knowledge Assets Knowledge is intangible and difficult to measure/value Most knowledge loses value or depreciates over time The value of knowledge depends on several factors: * user need: correct context and level of knowledge * user experience and awareness of value * formalization and organization of knowledge * knowledge currency: latest/best research and results * knowledge availability: ease of access and timeliness * format options and presentation customization * ease of sharing knowledge Knowledge provides organizational value through its creation, sharing & use: * improved performance * more effective management, better decisions * increased learning and innovation * increased knowledge sharing, teamwork, collaboration * increased value of knowledge asset through formalization & organization of knowledge * anticipated benefits: synergy & cooperation Knowledge also conveys personal/political power: * knowledge hoarding gives power to the holder who gain personally by filtering/ suppressing bad news and broadcasting good news * openness and sharing may force functions to act in ways opposed to local interests © 2006 Tom Beckman

Methods to Increase Intellectual Asset Value Collect/Uncover Existing Knowledge * read & research * Methods to Increase Intellectual Asset Value Collect/Uncover Existing Knowledge * read & research * attend conferences & seminars * buy commercial databases and knowledge sources Make Knowledge Explicit * put knowledge in electronic form * collect existing available explicit knowledge * query humans for implicit knowledge * observe behaviors, relationships and events * elicit tacit knowledge from domain experts Organize and Structure Explicit Knowledge * create a knowledge repository and Web site * classify, edit and maintain the repository Share Existing Knowledge * access knowledge repository and Web site * develop communities of practice * sponsor conference & workshops * build help desks Apply Knowledge to Perform Work Create New Knowledge * research, reflect, discuss, hypothesize, experiment Transform Knowledge Up the Knowledge Hierarchy © 2006 Tom Beckman

Technology Resource Class Smart System Methodology incorporates many innovative IT concepts, disciplines and technologies: Technology Resource Class Smart System Methodology incorporates many innovative IT concepts, disciplines and technologies: o Enterprise Architecture o Artificial intelligence, Advisory Systems, Object methods o Machine learning: deduction, induction, genetic algorithms, NN o Natural language and text generation o Knowledge representation & knowledge elicitation o Knowledge exploration & discovery * data mining * text mining Proper application of IT can enable all Business System Model components and greatly increase Stakeholder value * data visualization * link and event analysis o Knowledge repository & performance support systems o E-Learning and intelligent tutoring systems o Web services: portal, search, content management o Semantic technology: taxonomies, metatagging o Business and competitive intelligence © 2006 Tom Beckman

Intellectual Asset Resource Class Resource conservation does not apply to knowledge * knowledge is Intellectual Asset Resource Class Resource conservation does not apply to knowledge * knowledge is not consumed or depleted through use * knowledge is multiplied through sharing * knowledge often increases through sharing and use * knowledge can be used simultaneously by multiple processes Intellectual assets in human form are volatile and can disappear or become unavailable overnight Knowledge Management, a type of resource management * delivered at the right time * available at the right place * present in the right format * satisfies quality requirements * obtained at the lowest possible cost * employees are promoted or get new job assignments * employees leave the firm for other career opportunities * employees retire or are downsized/dismissed * functions are outsourced * employees/organizations may refuse to share expertise Costs of acquiring/developing/creating knowledge * knowledge acquisition costs can vary widely (from buying books, to knowledge elicitation and new hires) * development costs are often high (lessons learned, education, learning, research, experimenting) * costs of storing, sharing, using knowledge are low, especially if already have IT infrastructure © 2006 Tom Beckman

Building Competency and Knowledge « IRS Criminal Investigation Special Agents identify & develop fraud Building Competency and Knowledge « IRS Criminal Investigation Special Agents identify & develop fraud cases for prosecution by Justice Dept « The unit of work is the case; 18 months is average time to work 1. Novices are taught concepts and case exemplars in classroom by experts and senior practitioners 7. Senior practitioners and experts also develop & refine Knowledge Repositories, consisting of documents and Expert Systems 2. Novices apply concepts and examples to real cases under guidance of senior practitioners 3. Practitioners work cases and document their experiences & results from cases into a Case. Base (a type of expert system & a part of a knowledge repository) 4. Case-based reasoning expert systems are used to 6. Senior practitioners and research and find similar experts engage in Communities cases of Practice and hold peer workshops to further develop 5. Experts abstract and generalize from cases and create models their concepts, models and (Model-Based Reasoning), practices principles, rules (Rule-Based Systems) and guidelines © 2006 Tom Beckman

Knowledge Representations for Semantic Interoperability Questions? Tom Beckman Principal, Beckman Associates tombeckman@starpower. net 202 Knowledge Representations for Semantic Interoperability Questions? Tom Beckman Principal, Beckman Associates [email protected] net 202 -362 -5774 © 2006 Tom Beckman