Outline

• Semantics: an Overview
• The Logos Controlled Vocabulary (LCV)?
• Use Cases
• References for Concepts
• Concept "Importance"
• User Contribution (Logos Topics)

What Does "Semantics" Mean?

Data that has:

• Identity
  • Unambiguous
  • Addressable
• Utility
• Relationships

Semantics and Linked Data

Compare TimBL's rules for Linked Data:

1. Use URIs to identify things (= Identity)
   • Use HTTP URIs so people can look things up
2. Provide useful information in a standard format when someone references a URI (= Utility)
3. Include links to other URIs (= Relationships)

Semantic Organization at Logos

... organizing the world's Biblical information

• the Bible Knowledgebase: people, places, and things described in the Biblical text
• the Logos Controlled Vocabulary
• Bob Pritchett's talk at O'Reilly Tools of Change for Publishing Conference 2010: Domain-Specific Tools to Add Value to E-Books

Other Linked-Data Semantics in Logos

• ... and other examples
• Content and links not always formally represented

The Logos Controlled Vocabulary (LCV)

• LCV = a controlled vocabulary for Biblical studies
  • controlled vocabulary: an organized system for labeling content using words and phrases, to improve search and navigation
  • Mitigates problems of ambiguity, homographs, synonyms, variation in spelling and other kinds of representation
• Organizes content from Bible dictionaries, encyclopedias, and other subject-oriented secondary resources, regardless of how the publisher expressed it
• Provides semantic organization of concepts
  • Modeling the domain rather than annotating texts
  • Supporting the end-user rather than the philosopher

LCV Example: Prisca

• Example: %Prisca_Person/"Prisca"
• Utility:

  Type: Person	Name: True
  Preferred label: "Prisca"	Extra-biblical: False
  Alternate labels: "Priscilla"	Compound: False

• Relationships:

  BK Entities: #Prisca.1

  Articles: Anchor/PRISCAPERSON, TYNBIBDCT/L4559, ISBE/PRISCA, ...

  Logos Topics: http://topics.logos.com/Prisca

  Wikipedia: Priscilla and Aquila

LCV Example: Deceit

• Example: %deceit/"Deceit"
• Utility:

  Type: none	Name: False
  Preferred label: "Deceit"	Extra-biblical: False
  Alternate labels: "Deception", "Deceitful", "Deceive"	Compound: False

• Relationships:

  Articles: ISBE/DECEIT, NBD/R494, ...

  Logos Topics: http://topics.logos.com/Deceit

LCV Development

• Ambiguity: "communion" as Lord's supper, or as fellowship
• Homographs: "bear" as carrying, versus an animal, versus a constellation
• Variation: "Pilate" vs. "Pontius Pilate" vs. "Pilate, Pontius"

LCV Development

• Initial version developed by merging content from 7 Bible dictionaries
  • Extract headwords
  • Align automatically (but conservatively), merging when possible
  • Review manually
• Total initial terms: > 40k
• After automatic merging: ~23k
• Current inventory: ~11k concepts

LCV Development Continues

• Merging additional resources suggests new concepts:
  • Archaeological Dictionary of the Holy Land: 90/547 (16%)
    • Mostly very specific locations like %EinSamiya_Place
  • Nelson's Illustrated Bible Dictionary: 200/4833 (4%)
  • Harper's Bible Dictionary: 81/2962 (3%)
  • May be some overlap among these, particular for archaic KJV-based concepts
• Significant holes still remain:
  • Individuals from church history
  • More specialized theological concepts

LCV Types

Some concepts are still general: %Simon_Person refers to content about multiple Biblical people with that name (Compound: True)

Use Cases

• Improved "topic" search:
  • Link to the same concept regardless of how originally labeled
  • Provide consistent semantics for content
  • Suggest alternate concepts for the same term
• Support exploration by suggesting concepts for a passage

From Concepts to References

• Content about Biblical subjects typically includes references
• Very useful to provide the reverse index: given a passage, what concepts are associated with it?
• Estimate this by data mining and aggregating counts across all content and concepts:
  • Each article "votes" on the most likely references for the concept
  • Each concept "votes" on the most likely concepts for a reference
• Ideally, these estimates should get better as more content is added

Reference Aggregation for %Prisca Content

Reference Aggregation for %Gabriel

Challenges with Reference Counts

• How to combine counts
  • across overlapping references
  • for reference intersection with passage queries
• References might count less
  • if an article is longer
  • if the reference occurs later in an article
• Longer references are often more weakly "about" a concept
  • e.g. Is 40-66 for %Isaiah_Person
• We only know that a reference occurs, not why, or what it means
• We also manually provide "key references" for a small number of concepts

Reference-Based Retrieval: Acts 18.26

Reference-Based Retrieval: Luke 1.11-13

Concept "Importance"

• Concept content can be both broad (many sources) and deep (many words or references)
• Concepts can be popular (based on user activity)
• Concepts can be significant (by some objective standard)

Logos Topics

• Selected attributes for all concepts on the web
  • Preferred and alternate terms
  • Key passages
  • Wiki format (but more like dmoz than Wikipedia)
• Goal: engage user community in enriching data
  • Reading lists
  • Web links

User Contributions to Logos Topics

Site was launched in August 2009, with 10779 pages (one per concept). Since that time:

• Reading lists have been added to 118 pages, including:
  • Leviathan
  • Steward
• 259 additional pages have been created, including:
  • Book of Josippon
• > 1400 other page edits, including:
  • Aaron's Rod

Future Directions

• Definitions
• Hierarchy to support broadening/narrowing search
• Incorporate user contributions
• Linkage to WordNet and similar resources

Conclusions

• Described a controlled vocabulary for Biblical studies
  • Supports search and discovery for subject-oriented content
  • Shipping now, with aspects on the web
• A growing body of linked data for Biblical studies