Transcript Slide 1

High-level goals and low-level systems: how can
computers help solving complex problems?
Piek Vossen and Roxane Segers
VU University Amsterdam
KYOTO (ICT-211423)
Yielding Ontologies for Transition-Based Organization
FP7: Intelligent Content and Semantics
http://www.kyoto-project.eu/
1st Kyoto workshop:
Environmental Knowledge Transition and Exchange
February, 2-3, 2009, Artis Amsterdam
Overview
1. analysis of the user questions
(provided by ECNC and WWF)
a. clear cases
b. unclear cases and possible solutions
2. Concepts and relations
3. Discussion: how to use knowledge discovery to answer
complex questions
ICT-211423
1st Kyoto workshop: Environmental Knowledge Transition
and Exchange, February, 2-3, 2009, Artis Amsterdam
Analysis of the user questions
▪Total of 330 questions from WWF and
ECNC.
▪ Analysis of question type (what, where,
who, when, how etc).
▪ Words that express the relation that is
asked for
ICT-211423
1st Kyoto workshop: Environmental Knowledge Transition
and Exchange, February, 2-3, 2009, Artis Amsterdam
Classification of questions
•
Finding Definitions, e.g.
– What are the drivers?
– What are ecosystem services?
•
Pin-Pointing Sources
– Where are biodiversity hotspots?
– Who is an expert on biodiversity?
•
Information Retrieval, e.g.
– What is the current status of a specie?
– What are the most recent countings of a certain specie?
•
Examples (of best practices), e.g.
– Best practices of resource management in poor area's?
– Are there examples of nature conservation and poverty elivation?
– What are good examples of E.S.?
•
High level questions, e.g.
–
–
–
–
What are the key biodiversity indicators in a certain area?
What is the effect of climate change and species diversity?
What is the impact of declining bee populations on agricultural productivity?
Which rare and threatened species/ecosystems are people most dependent on
for economic benefits?
ICT-211423
1st Kyoto workshop: Environmental Knowledge Transition
and Exchange, February, 2-3, 2009, Artis Amsterdam
Baseline retrieval results
30 high-level questions,
Result
Rank
CONFIRMED
DISAPPROVED
UNDECIDED
Total
0
13
20.31%
27
20.30%
10
15.87%
50
19.23%
1
6
9.38%
9
6.77%
9
14.29%
24
9.23%
2
8
12.50%
13
9.77%
7
11.11%
28
10.77%
3
5
7.81%
6
4.51%
3
4.76%
14
5.38%
4
8
12.50%
6
4.51%
2
3.17%
16
6.15%
5
2
3.13%
7
5.26%
3
4.76%
12
4.62%
6
2
3.13%
6
4.51%
4
6.35%
12
4.62%
7
2
3.13%
2
1.50%
1
1.59%
5
1.92%
8
4
6.25%
3
2.26%
1
1.59%
8
3.08%
9
1
1.56%
5
3.76%
0
0.00%
6
2.31%
-1
13
20.31%
49
36.84%
23
36.51%
85
32.69%
64
24.62%
133
51.15%
63
24.23%
260
Total
ICT-211423
1st Kyoto workshop: Environmental Knowledge Transition
and Exchange, February, 2-3, 2009, Artis Amsterdam
Some numbers…
Example of question analysis:
• What is the effect of poverty on biodiversity?
• How can sustainable financing influence livelihood?
What is
How
What/Which
Auxiliary verbs
Where
Why
When
Who
ICT-211423
causes
causes
41.2%
23.9%
16.9%
9.3%
3.9%
1.2%
0.6%
0.6%
1st Kyoto workshop: Environmental Knowledge Transition
and Exchange, February, 2-3, 2009, Artis Amsterdam
Difficult causal relations
(no search results)
• What is the impact of dogs on wildlife?
Canine x Process (spreading disease,
hunting?)
Organism x Process (decreasing?)
• What is the effect of hedgerows on air
quality?
Artefact x Process
Attribute x Process (Pollution?)
ICT-211423
1st Kyoto workshop: Environmental Knowledge Transition
and Exchange, February, 2-3, 2009, Artis Amsterdam
Association
What is the effect of hedgerows on air quality?
AIR QUALITY
HEDGEROW
wildlife corridor
soil erosion
aerosols
biodiversity
aerosols
ICT-211423
air pollution
biodiversity
1st Kyoto workshop: Environmental Knowledge Transition
and Exchange, February, 2-3, 2009, Artis Amsterdam
Decomposition of a high level question
hedgerows
air quality
role
capture,
interception
aerosols
airborne
particles
hyponym
aerosols
ICT-211423
hyponym
dust
1st Kyoto workshop: Environmental Knowledge Transition
and Exchange, February, 2-3, 2009, Artis Amsterdam
smog
What does it take?
• Extract all events and properties involving
hedgerows ► direct relations, causal chains,
roles, etc.
• Extract all events and properties involving air
quality ► direct relations, causal chains, roles,
etc
• Do the same for any synonyms, hyponyms cohyponyms, or hyperonyms
• Extract dates and locations for events and
properties
• Try to intersect and quantify the results
ICT-211423
1st Kyoto workshop: Environmental Knowledge Transition
and Exchange, February, 2-3, 2009, Artis Amsterdam
Small preliminary domain model
Are there huge negative effects with regard to econetworks and alien invasive species?
Area
eco-network
core area
breeding pond
buffer area
[area connector]
stepping stone
Cultural
hedgerow
canal
stonewall
corridor
Natural
watercourses
rivers
coastlines
ICT-211423
migration corridor
commuting corridor
dispersal corridor
bird migration corridor
terrestrial dispersal corridor
1st Kyoto workshop: Environmental Knowledge Transition
and Exchange, February, 2-3, 2009, Artis Amsterdam
Versions of KYOTO
• KYOTO-1:
– extract all countings of species & populations
in regions for dates and periods
• KYOTO-2:
– changes in numbers of species and sizes of
populations, temporal relations and causal
relations
• KYOTO-3:
– to be determined....
ICT-211423
1st Kyoto workshop: Environmental Knowledge Transition
and Exchange, February, 2-3, 2009, Artis Amsterdam