|
1
|
- Steve Maslennikov and
- Chan Yee Seng
|
|
2
|
- Motivation
- Search Interfaces
- Name Entitie Extraction
- Combination of Search Interfaces
- Conclusion
|
|
3
|
- Search in classical Digital Library (DL): keyword, classification,
metadata
- What if we want to search for photographs, maps, major historical
events?
- Search DL using characterizing attribute
- Brings us to our topic…
|
|
4
|
- Digital library is a good source for information
- Problem: how to access?
- Classical IR models don’t work well with long documents and other types
of information (e.g. maps)
|
|
5
|
|
|
6
|
- Navigational Classification Hierarchy
- Digital Library for Earth System Education
- (DLESE)
|
|
7
|
|
|
8
|
|
|
9
|
|
|
10
|
|
|
11
|
- Informative Map-interface
- Perseus Project
|
|
12
|
|
|
13
|
|
|
14
|
- Navigational Map-interface
- Alexandria Digital Library (ADL)
|
|
15
|
|
|
16
|
|
|
17
|
|
|
18
|
|
|
19
|
|
|
20
|
- map, classification, time period
- Map provides navigation and specifies constraints
- Visual information about categories
- NE extraction adds new dimensions (DATE, LOC, NAME etc.)
|
|
21
|
- Search interface should be intuitive
- Visual techniques show the content of very long documents with different
types of information
|
|
22
|
- http://www.dlese.org/dds/index.jsp
- http://webclient.alexandria.ucsb.edu/
- http://www.perseus.tufts.edu/
|