Bergman A Knowledge Representation Practionary

Guidelines Based on Charles Sanders Peirce
1. Auflage 2018
ISBN: 978-3-319-98092-8
Verlag: Springer International Publishing
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, 462 Seiten

This major work on knowledge representation is based on the writings of Charles S. Peirce, a logician, scientist, and philosopher of the first rank at the beginning of the 20th century. This book follows Peirce's practical guidelines and universal categories in a structured approach to knowledge representation that captures differences in events, entities, relations, attributes, types, and concepts. Besides the ability to capture meaning and context, the Peircean approach is also well-suited to machine learning and knowledge-based artificial intelligence. Peirce is a founder of pragmatism, the uniquely American philosophy.Knowledge representation is shorthand for how to represent human symbolic information and knowledge to computers to solve complex questions. KR applications range from semantic technologies and knowledge management and machine learning to information integration, data interoperability, and natural language understanding. Knowledge representation is an essential foundation for knowledge-based AI.This book is structured into five parts. The first and last parts are bookends that first set the context and background and conclude with practical applications. The three main parts that are the meat of the approach first address the terminologies and grammar of knowledge representation, then building blocks for KR systems, and then design, build, test, and best practices in putting a system together. Throughout, the book refers to and leverages the open source KBpedia knowledge graph and its public knowledge bases, including Wikipedia and Wikidata. KBpedia is a ready baseline for users to bridge from and expand for their own domain needs and applications. It is built from the ground up to reflect Peircean principles.This book is one of timeless, practical guidelines for how to think about KR and to design knowledge management (KM) systems. The book is grounded bedrock for enterprise information and knowledge managers who are contemplating a new knowledge initiative.This book is an essential addition to theory and practice for KR and semantic technology and AI researchers and practitioners, who will benefit from Peirce's profound understanding of meaning and context.

Michael K. Bergman is a senior principal for Cognonto Corporation, and lead editor for the open-source KBpedia knowledge structure. For more than a decade, his AI3:::Adaptive Information blog has been a leading go-to resource on topics in semantic technologies, large-scale knowledge bases for machine learning, data interoperability, knowledge graphs and mapping, and fact and entity extraction and tagging. For the past twenty years Mike has been an entrepreneur, Web scientist, and independent consultant. For the decade up to 2018, Mike was the CEO of Structured Dynamics LLC, which he co-founded with Fred Giasson.  Mike has held C-class positions and was a founder of the prior companies Zitgist LLC, BrightPlanet Corporation, VisualMetrics Corporation, and TheWebTools Company. These companies provided notable market advances in semantic technologies, data warehousing, the deep Web, large-scale Internet databases, meta-search tools, and bioinformatics. Bergman began his professional career in the mid-1970s as a project director for the U.S. EPA for a major energy study called the Coal Technology Assessment. He later taught in the Graduate School of Engineering at the University of Virginia, where he was a fellow in the Energy Policies Study Center. He then joined the American Public Power Association in 1982, where he rose to director of energy research. APPA's pioneering work with small computers sparked Bergman's transition to information technologies. Before entering industry, Mike was a doctoral candidate at Duke University in population genetics.

Autoren/Hrsg.

Bergman, Michael K.

Inhaltsverzeichnis

1;Dedication;5
2;Preface;6
3;In-line Citations;12
4;Contents;13
5;Chapter 1: Introduction;16
5.1;Structure of the Book;17
5.2;Overview of Contents;18
5.3;Key Themes;25
5.4;References;28
6;Chapter 2: Information, Knowledge, Representation;29
6.1;What Is Information?;30
6.1.1;Some Basics of Information;30
6.1.2;The Structure of Information;33
6.1.2.1;Forms of Structure;33
6.1.2.2;Some Structures Are More Efficient;34
6.1.2.3;Evolution Favors Efficient Structures;35
6.1.3;The Meaning of Information;37
6.2;What Is Knowledge?;41
6.2.1;The Nature of Knowledge;41
6.2.2;Knowledge as Belief;44
6.2.3;Doubt as the Impetus of Knowledge;46
6.3;What Is Representation?;47
6.3.1;The Shadowy Object;48
6.3.2;Three Modes of Representation;50
6.3.3;Peirce’s Semiosis and Triadomany;52
6.4;References;55
7;Part I: Knowledge Representation in Context;57
7.1;Chapter 3: The Situation;58
7.1.1;Information and Economic Wealth;59
7.1.1.1;Historical Breakpoints;59
7.1.1.2;The X Factor of Information;63
7.1.1.3;Knowledge and Innovation;64
7.1.2;Untapped Information Assets;67
7.1.2.1;Valuing Information as an Asset;68
7.1.2.2;Lost Value in Information;70
7.1.2.3;The Information Enterprise;72
7.1.3;Impediments to Information Sharing;74
7.1.3.1;Cultural Factors;74
7.1.3.2;Tooling and Technology;75
7.1.3.3;Perspectives and Priorities;76
7.1.4;References;76
7.2;Chapter 4: The Opportunity;78
7.2.1;KM and a Spectrum of Applications;79
7.2.1.1;Some Premises;79
7.2.1.2;Potential Applications;80
7.2.1.3;A Minimal Scaffolding;81
7.2.2;Data Interoperability;82
7.2.2.1;The Data Federation Pyramid;82
7.2.2.2;Benefits from Interoperability;84
7.2.2.3;A Design for Interoperating;85
7.2.3;Knowledge-Based Artificial Intelligence;87
7.2.3.1;Machine Learning;92
7.2.3.2;Knowledge Supervision;94
7.2.3.3;Feature Engineering;96
7.2.4;References;97
7.3;Chapter 5: The Precepts;98
7.3.1;Equal Class Data Citizens;99
7.3.1.1;The Structural View;100
7.3.1.2;The Formats View;101
7.3.1.3;The Content View;102
7.3.2;Addressing Semantic Heterogeneity;104
7.3.2.1;Sources of Semantic Heterogeneity;104
7.3.2.2;Role of Semantic Technologies;108
7.3.2.3;Semantics and Graph Structures;110
7.3.3;Carving Nature at the Joints;110
7.3.3.1;Forming ‘Natural’ Classes;111
7.3.3.2;A Mindset for Categorization;115
7.3.3.3;Connections Create Graphs;116
7.3.4;References;117
8;Part II: A Grammar for Knowledge Representation;118
8.1;Chapter 6: The Universal Categories;119
8.1.1;A Foundational Mindset;120
8.1.1.1;A Common Grounding in Peirce;120
8.1.1.2;Truth Is Testable and Fallible;121
8.1.1.3;Upper Ontologies, Context, and Perspective;122
8.1.1.4;Being Attuned to Nature;123
8.1.2;Firstness, Secondness, Thirdness;124
8.1.2.1;Constant Themes of Three;124
8.1.2.2;Summary of the Universal Categories;125
8.1.2.3;The Irreducible Triad;127
8.1.3;The Lens of the Universal Categories;128
8.1.3.1;An Aha! Moment;129
8.1.3.2;Grokking the Universal Categories;130
8.1.3.3;Applying the Universal Categories;135
8.1.3.3.1;The Categories and Categorization;136
8.1.4;References;138
8.2;Chapter 7: A KR Terminology;140
8.2.1;Things of the World;142
8.2.1.1;Entities, Attributes, and Concepts;142
8.2.1.2;What Is an Event?;144
8.2.2;Hierarchies in Knowledge Representation;146
8.2.2.1;Types of Hierarchical Relationships;147
8.2.2.2;Structures Arising from Hierarchies;152
8.2.3;A Three-Relation Model;154
8.2.3.1;Attributes, the Firstness of Relations;156
8.2.3.2;External Relations, the Secondness of Relations;157
8.2.3.3;Representations, the Thirdness of Relations;157
8.2.3.4;The Basic Statement;158
8.2.4;References;159
8.3;Chapter 8: KR Vocabulary and Languages;161
8.3.1;Logical Considerations;163
8.3.1.1;First-Order Logic and Inferencing;164
8.3.1.1.1;Deductive Logic;166
8.3.1.1.2;Inductive Logic;167
8.3.1.1.3;Abductive Logic;168
8.3.1.2;Redux: The Nature of Knowledge;170
8.3.1.3;Particulars, Generals, and Description Logics;172
8.3.2;Pragmatic Model and Language Choices;173
8.3.2.1;RDF: A Universal Solvent;173
8.3.2.2;OWL 2: The Knowledge Graph Language;175
8.3.2.3;W3C: Source for Other Standards;176
8.3.3;The KBpedia Vocabulary;177
8.3.3.1;Structured on the Universal Categories;177
8.3.3.2;Three Main Hierarchies;178
8.3.3.2.1;The Instances Vocabulary;178
8.3.3.2.2;The Relations Vocabulary;180
8.3.3.2.2.1;Attributes Relations (1ns);184
8.3.3.2.2.2;External Relations (2ns);184
8.3.3.2.2.3;Representation Relations (3ns);185
8.3.3.2.3;The Generals (KR Domain) Vocabulary;186
8.3.3.3;Other Vocabulary Considerations;187
8.3.4;References;190
9;Part III: Components of Knowledge Representation;191
9.1;Chapter 9: Keeping the Design Open;192
9.1.1;The Context of Openness;193
9.1.1.1;An Era of Openness;194
9.1.1.2;The Open-World Assumption;198
9.1.1.3;Open Standards;201
9.1.2;Information Management Concepts;202
9.1.2.1;Things, Not Strings;203
9.1.2.2;The Idea and Role of Reference Concepts;204
9.1.2.3;Punning for Instances and Classes;208
9.1.3;Taming a Bestiary of Data Structs;209
9.1.3.1;Rationale for a Canonical Model;210
9.1.3.2;The RDF Canonical Data Model;211
9.1.3.3;Other Benefits from a Canonical Model;213
9.1.4;References;213
9.2;Chapter 10: Modular, Expandable Typologies;215
9.2.1;Types as Organizing Constructs;216
9.2.1.1;The Type-Token Distinction;216
9.2.1.2;Types and Natural Classes;218
9.2.1.3;Very-Fine-Grained Entity Types;220
9.2.2;A Flexible Typology Design;223
9.2.2.1;Construction of the Hierarchical Typologies;223
9.2.2.2;Typologies Are Modular;224
9.2.2.3;Typologies Are Expandable;226
9.2.3;KBpedia’s Typologies;227
9.2.3.1;Full Listing of Typologies;227
9.2.3.2;‘Core’ Typologies;230
9.2.3.3;Tailoring Your Own Typologies;234
9.2.4;References;234
9.3;Chapter 11: Knowledge Graphs and Bases;235
9.3.1;Graphs and Connectivity;236
9.3.1.1;Graph Theory;237
9.3.1.2;The Value of Connecting Information;239
9.3.1.3;Graphs as Knowledge Representations;244
9.3.2;Upper, Domain, and Administrative Ontologies;245
9.3.2.1;A Lay Introduction to Ontologies;246
9.3.2.2;Ontologies Are a Family of Graphs;247
9.3.2.2.1;Incipient Potentials;248
9.3.2.3;Good Ontology Design and Construction;249
9.3.3;KBpedia’s Knowledge Bases;250
9.3.3.1;KBpedia KBs;251
9.3.3.1.1;Primary KBs;251
9.3.3.1.2;Secondary KBs;253
9.3.3.2;Candidate KBs for Expansion;254
9.3.4;References;254
10;Part IV: Building KR Systems;256
10.1;Chapter 12: Platforms and Knowledge Management;257
10.1.1;Uses and Work Splits;258
10.1.1.1;The State of Tooling;258
10.1.1.2;TBox, ABox, and Work Splits;260
10.1.1.3;Content Workflows;265
10.1.2;Platform Considerations;268
10.1.2.1;Supporting Multiple Purposes;269
10.1.2.1.1;Search;269
10.1.2.1.2;Knowledge Management;270
10.1.2.2;An Ontology-Based Design;271
10.1.2.3;Enterprise Considerations;272
10.1.3;A Web-Oriented Architecture;274
10.1.3.1;Web Orientation and Standards;275
10.1.3.2;A Modular Web Service Design;275
10.1.3.3;An Interoperability Architecture;277
10.1.4;References;278
10.2;Chapter 13: Building Out the System;279
10.2.1;Tailoring for Domain Uses;280
10.2.1.1;A Ten-Point Checklist for Domain Use;280
10.2.1.2;An Inventory of Assets;281
10.2.1.3;Phased Implementation Tasks and Plan;282
10.2.1.3.1;Domain Knowledge Graph;283
10.2.1.3.2;Instance Data Population;284
10.2.1.3.3;Analysis and Content Processing;284
10.2.1.3.4;Use and Maintenance;285
10.2.1.3.5;Testing and Mapping;285
10.2.1.3.6;Documentation;286
10.2.2;Mapping Schema and Knowledge Bases;286
10.2.2.1;Mapping Methods and Tools;286
10.2.2.2;Building Out the Schema;287
10.2.2.2.1;Overview of Approaches;288
10.2.2.2.2;Some Design Guidelines;290
10.2.2.2.2.1;Be Lightweight and Modular;290
10.2.2.2.2.2;Use Reference Structures;291
10.2.2.2.2.3;Reuse Existing Structure;292
10.2.2.2.2.4;Build Incrementally;292
10.2.2.2.2.5;Use Simple Predicates;293
10.2.2.2.2.6;Test for Logic and Consistency;293
10.2.2.2.2.7;Map to External Ontologies;294
10.2.2.3;Building Out the Instances (Knowledge Bases);294
10.2.2.3.1;Update Changing Knowledge;295
10.2.2.3.2;Process the Input KBs;295
10.2.2.3.3;Install, Run, and Update the System;295
10.2.2.3.4;Test and Vet Placements;295
10.2.2.3.5;Test and Vet Mappings;296
10.2.2.3.6;Test and Vet Assertions;296
10.2.2.3.7;Ensure Completeness;296
10.2.2.3.8;Test and Vet Coherence;296
10.2.2.3.9;Generate Training Sets;296
10.2.2.3.10;Test and Vet Learners;297
10.2.2.3.10.1;Rinse and Repeat;297
10.2.3;Pay as You Benefit;297
10.2.3.1;Placing the First Stake;298
10.2.3.2;Incremental Build-Outs Follow Benefits;298
10.2.3.3;Learn to Quantify and Document Benefits;299
10.2.4;References;299
10.3;Chapter 14: Testing and Best Practices;301
10.3.1;A Primer on Knowledge Statistics;302
10.3.1.1;Two Essential Metrics, Four Possible Values;302
10.3.1.2;Many Useful Statistics;305
10.3.1.3;Working Toward ‘Gold Standards’;307
10.3.2;Builds and Testing;310
10.3.2.1;Build Scripts;311
10.3.2.2;Testing Scripts;312
10.3.2.3;Literate Programming;313
10.3.3;Some Best Practices;315
10.3.3.1;Data and Dataset Practices;316
10.3.3.1.1;Dataset Best Practices;316
10.3.3.1.2;Linked Data;317
10.3.3.2;Knowledge Structures and Management Practices;318
10.3.3.2.1;Organizational and Collaborative Best Practices;318
10.3.3.2.2;Naming and Vocabulary Best Practices;318
10.3.3.2.3;Best Ontology Practices;319
10.3.3.3;Testing, Analysis, and Documentation Practices;320
10.3.3.3.1;Testing Best Practices;320
10.3.3.3.2;Analytical Best Practices;320
10.3.3.3.3;Documentation Best Practices;321
10.3.4;References;322
11;Part V: Practical Potentials and Outcomes;323
11.1;Chapter 15: Potential Uses in Breadth;324
11.1.1;Near-Term Potentials;325
11.1.1.1;Word Sense Disambiguation;325
11.1.1.2;Relation Extraction;327
11.1.1.3;Reciprocal Mapping;328
11.1.1.4;Extreme Knowledge Supervision;330
11.1.2;Logic and Representation;332
11.1.2.1;Automatic Hypothesis Generation;332
11.1.2.2;Encapsulating KBpedia for Deep Learning;334
11.1.2.3;Measuring Classifier Performance;335
11.1.2.4;Thermodynamics of Representation;336
11.1.3;Potential Methods and Applications;337
11.1.3.1;Self-Service Business Intelligence;337
11.1.3.2;Semantic Learning;338
11.1.3.3;Nature as an Information Processor;340
11.1.3.4;Gaia Hypothesis Test;342
11.1.4;References;344
11.2;Chapter 16: Potential Uses in Depth;347
11.2.1;Workflows and BPM;347
11.2.1.1;Concepts and Definitions;349
11.2.1.2;The BPM Process;350
11.2.1.3;Optimal Approaches and Outcomes;351
11.2.2;Semantic Parsing;353
11.2.2.1;A Taxonomy of Grammars;354
11.2.2.2;Computational Semantics;357
11.2.2.3;Three Possible Contributions Based on Peirce;358
11.2.2.3.1;Peircean POS Tagging;359
11.2.2.3.2;Machine Learning Understanding Based on Peirce;362
11.2.2.3.3;Peirce Grammar;363
11.2.3;Cognitive Robotics and Agents;365
11.2.3.1;Lights, Camera, Action!;366
11.2.3.2;Grounding Robots in Reality;369
11.2.3.3;Robot as Pragmatist;370
11.2.4;References;371
11.3;Chapter 17: Conclusion;374
11.3.1;The Sign and Information Theoretics;375
11.3.2;Peirce: The Philosopher of KR;376
11.3.2.1;Knowledge and Peirce;377
11.3.2.2;Time to Move from Theory to Practice;378
11.3.3;Reasons to Question Premises;380
11.3.3.1;AI Is a Field of KR;380
11.3.3.2;Hurdles to Be Overcome;381
11.3.3.3;Of Crystals and Robots;382
11.3.4;References;383
12;Appendix A: Perspectives on Peirce;384
12.1;Peirce, the Person;385
12.2;Peirce, the Philosopher;388
12.2.1;Peirce’s Architectonic;388
12.2.2;Chance, Existents, and Continuity: Real;390
12.2.2.1;Chance;391
12.2.2.2;Existents;392
12.2.2.3;Continuity;393
12.2.2.4;What Is Real;395
12.2.3;Leaning into Pragmatism;395
12.3;Peirce, the Polymath;396
12.3.1;Mathematics;397
12.3.2;Cenoscopy;398
12.3.3;Idioscopy;398
12.3.3.1;Scientist;399
12.3.3.2;Inventor;399
12.3.3.3;Humanist, as Person;400
12.4;An Obsession with Terminology;401
12.5;Peirce, the Polestar;402
12.6;Resources About Peirce;404
12.7;References;409
13;Appendix B: The KBpedia Resource;411
13.1;Components;412
13.1.1;The KBpedia Knowledge Ontology (KKO);413
13.1.2;The KBpedia Knowledge Bases;413
13.1.3;The KBpedia Typologies;415
13.2;Structure;416
13.3;Capabilities and Uses;420
14;Appendix C: KBpedia Feature Possibilities;422
14.1;What Is a Feature?;423
14.2;A (Partial) Inventory of Natural Language and KB Features;424
14.3;Feature Engineering for Practical Limits;432
14.4;Considerations for a Feature Science;433
14.5;Role of a Platform;434
15;Glossary;436
16;Index;451