Emergent Semantics

by Cudre-Mauroux; Philippe

• ISBN13: 9781420092271
• ISBN10: 1420092278
• Format: Hardcover
• Copyright: 2008-10-09
• Publisher: EFPL Press

Summary

Peer-to-peer systems are evolving with new information-system architectures, leading to the idea that the principles of decentralization and self-organization will offer new approaches in informatics, especially for systems that scale with the number of users or for which central authorities do not prevail. In this book, the author describes a new way of building global agreements, i.e. semantic interoperability, based only on decentralized, self-organizing interactions.

Table of Contents

Preface	p. xi
Introduction	p. 1
On Syntax, Semantics and Syntactic Semantics	p. 2
Emergent Semantics in Distributed Information Systems	p. 4
Scope of the Research	p. 5
What this Book is not about	p. 6
Outline	p. 7
Contributions	p. 9
On Integrating Data in the Internet Era	p. 13
Federated Databases	p. 13
XML, RDF and the Semantic Web	p. 17
Peer-to-Peer Information Management	p. 21
From unstructured to structured P2P Systems	p. 22
Peer Data Management	p. 24
Semantic Gossiping	p. 31
On Uncertain Schema Mappings in Decentralized Settings	p. 32
Mapping Completeness	p. 32
Mapping Soundness	p. 33
The Model	p. 34
The Data Model	p. 34
The Network Model	p. 35
Overview	p. 39
Syntactic Similarity	p. 41
Semantic Similarity	p. 45
Cycle Analysis	p. 45
Result Analysis	p. 52
Gossiping Algorithm	p. 55
Case Study	p. 58
Related Work	p. 60
Conclusions	p. 61
Self-Repairing Semantic Networks	p. 63
Experimental setup	p. 63
Cycle Analysis	p. 65
Result Analysis	p. 69
Combined Analysis	p. 72
Related Work	p. 73
Conclusions	p. 74
Probabilistic Message Passing	p. 75
Introduction	p. 75
Problem Definition	p. 76
An Introductory Example	p. 78
Modeling PDMSs as Factor-Graphs	p. 79
A Quick Reminder on Factor-Graphs and Message Passing Schemes	p. 80
On Factor-Graphs in Undirected PDMSs	p. 81
On Factor-Graphs in Directed PDMSs with Containment Mappings	p. 85
Embedded Message Passing	p. 88
On Feedback Variables in PDMS Factor-Graphs	p. 89
On Cycles in PDMS Factor-Graphs	p. 89
Embedded Message Passing Schedules	p. 91
Prior Belief Updates	p. 93
Introductory Example Revisited	p. 94
Performance Evaluation	p. 95
Performance Analyses	p. 96
Performance Evaluation on Random PDMS Networks	p. 99
Applying Message Passing on Real-World Schemas	p. 102
Conclusions	p. 105
Analyzing Semantic Interoperability in the Large	p. 107
Introduction	p. 107
The Model	p. 108
The Peer-to-Peer Model	p. 109
The Peer-to-Schema Model	p. 109
The Schema-to-Schema Model	p. 110
Semantic Interoperability In the Large	p. 111
Semantic Connectivity	p. 112
A Necessary Condition for Semantic Interoperability	p. 114
Undirected Model	p. 114
Directed Model	p. 116
Semantic Component Size	p. 120
Weighted Graphs	p. 121
Connectivity Indicator	p. 122
Giant Component Size	p. 123
Semantic Interoperability in a Bioinformatic Database Network	p. 124
The Sequence Retrieval System (SRS)	p. 125
Graph analysis of an SRS repository	p. 125
Applying the Heuristics to the SRS Graph	p. 127
Generating a Graph with a given Power-Law Degree Distribution	p. 127
Use Case Scenarios	p. 130
Conclusions	p. 134
GridVine: Building Internet-Scale Semantic Overlay Networks	p. 139
Introduction	p. 139
Overview of our Approach	p. 141
Data Independence	p. 141
Decentralized Semantics	p. 142
The P-Grid P2P System	p. 143
Semantic Support	p. 144
Metadata Storage	p. 145
Schema Definition And Storage	p. 146
Resolving Queries in GridVine	p. 148
Resolving Atomic Queries	p. 148
Resolving Conjunctive Queries	p. 150
Semantic Interoperability	p. 150
Schema Inheritance	p. 151
Semantic Gossiping	p. 152
Implementation	p. 154
Architectural Overview	p. 154
Querying	p. 155
Query Reformulation	p. 156
Experimental Evaluation	p. 157
Related Work	p. 159
Conclusions	p. 160
PicShark: Sharing Semi-Structured Annotations in the Large	p. 161
Introduction	p. 162
Sharing Semi-Structured Metadata	p. 163
On Semi-Structured Metadata	p. 163
On the Difficulty of Sharing Semi-Structured Metadata	p. 164
Opportunities for Reducing Metadata Scarcity Collaboratively	p. 165
Formal Model	p. 168
Metadata Entropy	p. 170
Recontextualizing Semi-Structured Metadata	p. 172
Exporting Local Metadata through Data Indexing	p. 173
Dealing with Metadata Incompleteness through Intra-Community Metadata Imputation	p. 174
Dealing with Metadata Heterogeneity through Pairwise Schema Mappings	p. 177
Dealing with Metadata Incompleteness through Inter-Community Metadata Propagation	p. 179
Possible Answers and User Feedback	p. 180
PicShark: Sharing Annotated Pictures in the Large	p. 180
Information Extraction in PicShark	p. 182
Performance Evaluation	p. 183
Related Work	p. 188
Conclusions	p. 190
idMesh: Graph-Based Disambiguation of Online Identities	p. 193
Introduction	p. 193
Contributions and Outline	p. 196
Related Work	p. 198
Problem Definition	p. 200
idMesh Constructs	p. 201
Making Sense of It	p. 203
An Introductory Example	p. 203
Deriving a Factor-Graph to Retrieve Equivalent Identities	p. 205
Deriving a Factor-Graph to Retrieve Up-to-date Identities	p. 210
Query Answering	p. 211
System Perspective	p. 211
Architectural Overview	p. 212
Distributed Probabilistic Inference	p. 213
Performance Evaluation	p. 214
Performance of the Inference Network	p. 215
Scale-Up	p. 218
Conclusions	p. 219
Conclusions	p. 221
List of Frequently Used Symbols and Abbreviations	p. 225
Bibliography	p. 227
Index	p. 243
Table of Contents provided by Ingram. All Rights Reserved.