A Web of Concepts

1
A Web of Concepts

• Dalvi, et al.
• Presented by Andrew Zitzelberger

2
Vision

• Transform hyperlinked bags of words into
  semantically rich aggregate view of information
  on the web.

3
Concept

• Things of interest
• Searching for information
• Accomplishing a task
• Reservations, etc.

4
Instances

• Record of a concept
• Restaurant
• Gochi (19980 Homestead Rd Cupertino CA)
• Academia?
• Publications, research institutions

5
Instance Representation

• Loosely-structured record (lrec)
• Attribute-key, value pairs
• Unique id field
• Entity matching problem
• Metadata
• Attribute list

6
Domain

• Set of related concepts
• Academic community domain publications,
  people, conferences

7
Usage StudyInstance vs. Concept Search

• yelp.com
• Month of queries resulting in a click
  (restaurants)
• 59 specific business URL
• 19 search URL either specific business or group
• 11 specific group URL

8
Usage StudyConcept Attribute Search

• Remove restaurant name and location information
  from query
• Co-occuring words
• Menu (3), coupons (1.8), online, weekly
  specials, locations (1.5)
• Nutrition, to go, delivery, careers, cod

9
Usage StudyAggregation Value

• 59 clicked on at least one other URL
• 35 clicked on at least two other URLs
• Small manual evaluation indicates pages are often
  about the same business.

10
Usage StudyConcepts vs. Browsing

• 42 of homepage visits are from search engine
• Immediately following URL
• 11.5 location
• 9 menu
• 1 coupons
• 10.5 of user trails contain more than one
  distinct instance of the restaurant concept

11
Extraction

• Create new records from the web
• Information extraction
• Linking
• Analysis
• Meta-data tagging (cuisine type)

12
Domain-centric vs. Site-centric Extraction

• Site-centric extraction
• Wrappers for page structure
• Probabilistic models (CRF)
• Domain-centric extraction
• Fields of interest
• Statistical properties (single zip code, etc.)
• Structure components (lists, link relationships)

13
Domain-centric Extraction

• Aggregator mining
• Learn from extracted knowledge (similar menus)
• Matching
• Text is about a record (restaurant review)

14
ApplicationAggregation
15
ApplicationSession Optimization

• User understanding
• Historical modeling
• Session modeling
• Content understanding
• Example Birks
• Birks and Mayors (luxury Jewelers) vs. Birks
  Steakhouse

16
ApplicationBrowse Optimization

• Alternatives (Restaurants)
• Similar type of cuisine
• Similar location
• Similar quality
• Augmentations (Camera)
• Batteries
• Memory cards

17
Concept Search
Result Pages shows multiple records Concept
Pages information about an instance Article
Pages a piece of authored text
18
Advertising

• Increase in targeted advertisements
• Target concepts rather than keywords

19
Challenges

• Transfer learning
• Transfer extractor knowledge
• Tracking uncertainty
• Accuracy issues
• Web of concepts is not a one time affair
• Wrapper problems
• Concept updates
• Relevance Measures
• User satisfaction

20
Related Work

• Information Extraction/Integration Systems
• Dataspace Systems
• Semantic Web

21
Future Work

• Enrich representation model
• Path storage to data
• Provenance, versions, uncertainty
• Hierarchal relationships (containment or
  inheritance)
• Ranking of disparate sources