Information management in the event of a crisis/disaster is a challenging problem as often the information is incomplete or inaccurate, while the public and the government both need to access to critical information in order to plan for crisis prevention and disaster relief. In addition, as communication networks and their bandwidth can be rather limited or overloaded in such events, it can be very difficult in aggregating, summarizing, and disseminating the relevant information in timely and useful ways. An ontology for the representation, processing, and integration of event information would be very useful in the application domain of crisis prevention and disaster relief. We have surveyed several event ontologies, and will propose some guidelines for the design of an event ontology for crisis-disaster information management.
An Event Ontology An Event Ontology An Event Ontology An Event Ontology An Event Ontology for Crisis for Crisis-Disaster Information Disaster Information The PNC 2010 Annual Conference and Joint Meetings The PNC 2010 Annual Conference and Joint Meetings The PNC 2010 Annual Conference and Joint Meetings The PNC 2010 Annual Conference and Joint Meetings December 1 December 1-3, 2010, 3, 2010, City University of Hong Kong City University of Hong Kong Hong Kong Hong Kong Andrea Wei Andrea Wei-Ching Huang and Tyng Ching Huang and Tyng-Ruey Chuang Ruey Chuang Institute of Information Science, Academia Sinica Institute of Information Science, Academia Sinica Ti iTi Ti iTi T aipei, T aiwan T aipei, T aiwan
Transcript
An Event Ontology
An Event OntologyAn Event OntologyAn Event Ontology An Event Ontology for Crisisfor Crisis--Disaster InformationDisaster Information
The PNC 2010 Annual Conference and Joint MeetingsThe PNC 2010 Annual Conference and Joint MeetingsThe PNC 2010 Annual Conference and Joint MeetingsThe PNC 2010 Annual Conference and Joint MeetingsDecember 1December 1--3, 2010, 3, 2010, City University of Hong Kong City University of Hong Kong
Hong Kong Hong Kong
Andrea WeiAndrea Wei--Ching Huang and TyngChing Huang and Tyng--Ruey Chuang Ruey Chuang Institute of Information Science, Academia SinicaInstitute of Information Science, Academia Sinica
M iM iM iM i R l tiR l tiR l tiR l ti Si li itSi li itSi li itSi li itMeaningMeaningMeaningMeaning RelationRelationRelationRelation SimplicitySimplicitySimplicitySimplicity
Huang & Chuang, 20104
An Event Ontology
“The world looks at a disasterdisaster,The world looks at a disasterdisaster,
often willing to help, but needing
an accurate picturean accurate picture. …
Humanity’s effectiveness Humanity’s effectiveness will be
WhyWhy do we need an event ontology for the disaster information management ?the disaster information management ?
Event Disaster Information Ontology
Event or Event concept is about dynamic change. It provides an effective way to
Disaster information has the characterizations of rapid changeability, ambiguity, vast
Ontology clarifies the relationship between objects stored in machines and provides
filter, share, communicate, organize, and integrate diverse and distributed information
quantity and diverse domains on the Web.
Disaster management
meanings both to human and machines work in coordination.
Existing event based ontologiesinformation.
Event descriptions usually focus on temporal and causal
Disaster management systems should not be isolated.
Traditional GIS system can
Existing event based ontologies are well equipped with spatial-temporal representations.
prelationships.
Traditional GIS system can not answer why question. W3C EIIF XG has worked on
the Who-What-Where, but no When-Why effort yet.The international Disaster Database (EM-DAT) has added
the ‘Event” on the top level for its data entry methodology
Huang & Chuang, 2010
the Event on the top level for its data entry methodology.
6
An Event Ontology
Li it tiLimitationsW h f l d fi i i li i i l iWe have no formal definition , linguistic analysis or philosophical exploration on “Event”; on “ontology” terms such as “data model”, “schemas” “classification”, “taxonomy”, or “thesaurus”; or on difference between “crisis”, “disaster”, “risk” or “emergency”.
We do not create a new event ontology for disaster management, instead we generalize a common Relationbased on the Event Concept as 5 General Event Relationsbased on the Event Concept, as 5 General Event Relations.
We have not implement or practice on this five GERs yet, and have not set logical axioms for their semantics Howeverand have not set logical axioms for their semantics. However our team is on the process of utilizing event concept to manage narrative accounts.
Huang & Chuang, 20107
An Event Ontology
G l & R ltGoal & Result
Huang & Chuang, 20108
An Event Ontology
EVENTEVENTEVENTEVENT
MeaningMeaningMeaningMeaning RelationRelationRelationRelation SimplicitySimplicitySimplicitySimplicitygggg p yp yp yp y
Huang & Chuang, 20109
An Event Ontology
2. Crisis-disaster Information
Miti tiMiti tiMiti tiMiti ti P dP dP dP d RRRR RRRRMitigationMitigationMitigationMitigation PreparednessPreparednessPreparednessPreparedness ResponseResponseResponseResponse RecoveryRecoveryRecoveryRecovery
Huang & Chuang, 201010
An Event Ontology
People need information as much as water, food, di i h ltmedicine, or shelter.
Information can save lives, livelihoods, and resources. It may be the only form of disaster preparedness that the most vulnerable can afford.
The right kind of information leads to a deeper understanding of needs and ways to respond.
The wrong information can lead to inappropriate, even dangerous interventions''
International Federation of the Red Cross and Red Crescent, World Disasters Report 2005
I H i tidi t l ( 2010) ‘In Hristidis et al. ( 2010) ‘s survey on data management & analysis in disaster situations, they suggest that:they suggest that:
employing ontology and semantic web technologies p y g gy gcan be used to identify and associate
ti ll di tsemantically corresponding concepts in the disaster-related information, so that the heterogeneous data can be integrated and ingestedheterogeneous data can be integrated and ingested.
Huang & Chuang, 201014
An Event Ontology
W3C Emergency InformationW3C Emergency Information Interoperability Frameworks
Huang & Chuang, 201015
DOLCE Lite basedWHO WHAT WHERE
An Event Ontology
W3C Emergency InformationW3C Emergency Information Interoperability Frameworks
Huang & Chuang, 201016
DOLCE Lite based
An Event Ontology
3. Semantic Web
Huang & Chuang, 201017
An Event Ontology
Computers used to be linkedlinked by telephone wires, by Internet, by Web of Documents, and then by Web of DataWeb of Data. .
The existence of linkslinksThe existence of links links themselves do not
carry meaningmeaningcarry meaning. meaning.
Huang & Chuang, 2010 Revised from Tim Berners-Lee , 2007/2010: http://www.w3.org/DesignIssues/Abstractions.html
18
An Event Ontology
Semantic WebSemantic Web: meaning not just spelling
“The Semantic Web is an extension of the current web in which information is given well-defined meaning, better enabling computers and people to work in cooperation ”enabling computers and people to work in cooperation.
http://www.w3.org/2001/sw/EO/points
OntologyOntology: machine-readable knowledge representationgygy g p
“a machine-understandable theory of (human) meaning.(C Legg 2007)(C.Legg, 2007)
A light-weight ontology example:RDF and RDF Schema
Graph with SemanticsGraph with Semantics1. Represent information.2. Express general knowledge.3. Implicit knowledge can be deduced
or inferred.
Huang & Chuang, 201021
An Event Ontology
Draw conclusions from the Given InformationGraph with SemanticsDraw conclusions from the Given Information.
Sebastian eats vegetable Thai g
Curry, …
th f h i therefore he is pitiable…
from Hitzler, Krotzsch and Rudolph (2009)
Huang & Chuang, 2010
22
An Event Ontology
3. Implicit knowledge is d d d i f ddeduced or inferred.
2. express more generic knowledgeknowledge
1. Represent Information
from Hitzler, Krotzsch and Rudolph (2009)
Fig 2.13: Graph representation of a simple RDFS ontologyHuang & Chuang, 2010
23
An Event Ontology
4. Event based Ontologiesg
R l tiR l tiR l tiR l tiRelationRelationRelationRelation
E tE tE tE tEventEventEventEvent
Huang & Chuang, 201024
An Event Ontology
What event based ontology can do?
for example, some of the general kinds of queries supported by Geospatial Event Model (GEM) include:
• What are all the events related to object X?
Wh t th bj t th t l t d t t Y?• What are the objects that are related to event Y?
• Can event Y happen without object X?
• What are all the events that are related to event Y?• What are all the events that are related to event Y?
• What events serve as initiator events for event Y?
• How many objects serve as event‐initiating or facilitating objects?• How many objects serve as event‐initiating or facilitating objects?
• What is the spatio‐temporal setting for event Y?
Huang & Chuang, 201025
Worboys and Hornsby ( 2004)
An Event Ontology
Event definitions in Stanford Encyclopedia of Philosophy y p p y
26
An Event Ontology
Different Classes with different hierarchy
4 core classes3 classes
17 properties1 class
6 properties27 class
6 perspectives3 classes1 function
5 classes6 functions13 relations
1 other class5 types
3 property constrains
5 classes20 properties
DUL classes6 DnS patterns1 other class
Huang & Chuang, 201027
An Event Ontology
Event definitions among different ontologies
SUMO: something that happens at a given place and time.DUL: Any physical social or mental process event or state
Event definitions among different ontologies
DUL: Any physical, social, or mental process, event, or state.BFO: similar to the " Temporal boundary of process" concept, which has been
defined as "a processual entity that is the fiat or bona fide instantaneous temporalboundary of a process.
1. EO: An arbitrary classification of a space/time region, by a cognitive agent. An event may have actively participating agents, passive factors, products, and a location in y y p p g g p pspace/time.
2. E: consider a Web in which each node is an event.3. GEM: event as process.4 UEO: entities in formal ontology as occurrence of actions & changes in the real world4. UEO: entities in formal ontology as occurrence of actions & changes in the real world.5. EM: event as actor and action in a situation ontology.6. F: " …perduring entities ( or perdurants or occurants ) that unfold over time, i.e., they
take up time..”7 LODE7. LODE: People conventionally refer to an action or occurrence taking place at a certain
time at a specific location as an event. Things that have happened or that are scheduled to happen.
8. SEM: Event cannot commit to a specific definition. Events encompass everything that
“Many Knowledge Representation systemsMany Knowledge Representation systems had a problem merging or interrelating two separate knowledge bases, as the model wasseparate knowledge bases, as the model was that any concept
had one and only one place had one and only one place in a tree of knowledge.
They therefore did not scale, or pass the test of independent invention ”test of independent invention.
http://www.w3.org/DesignIssues/RDFnot.html
Huang & Chuang, 201032
An Event Ontology
5. General Event Relation (GER)
Si li itSi li itSi li itSi li itR l tiR l tiR l tiR l ti
Most upper ontologies define classes andclasses and their tree of hierarchies.
By contrast, we generalize and define Event Relation mostly fromdefine Event Relation mostly from the existing Event based Ontologies.
Huang & Chuang, 201034
An Event Ontology
EVENT EVENT CONCEPTCONCEPTCONCEPTCONCEPT
SomethingSomething--AA SomethingSomething--BBSomethingSomething AA SomethingSomething BB
i di id l bj d ex individuals resource objects dataex. individuals, resource, objects, data bases, classes, domain ontology, upper ontology, framework, thesaurus, f d l
SimplicitySimplicitySimplicitySimplicityable to access key information concerning
disaster situation & resource availability.
Huang & Chuang, 201040
An Event Ontology
DisasterInfoDisasterInfo--BBDisasterInfoDisasterInfo--AA DisasterInfoDisasterInfo BBDisasterInfoDisasterInfo AA
Huang & Chuang, 2010
41
An Event Ontology
For mostFor most crisis/disaster/emergency information,
they have their domain categories/they have their domain categories/ hierarchies/ ontologies.
We want to We want to reuse them and provide freedom ofreuse them and provide freedom ofreuse them, and provide freedom of reuse them, and provide freedom of
meaning choices to users.meaning choices to users.
GER provides Event Concept to associate disaster information and relate to otherdisaster information and relate to other
useful resources according to users’ specific applications.
Huang & Chuang, 2010
specific applications. 42
An Event Ontology
EVENT EVENT CONCEPTCONCEPTCONCEPTCONCEPT
DisasterInfoDisasterInfo--AA DisasterInfoDisasterInfo--BBDisasterInfoDisasterInfo AA DisasterInfoDisasterInfo BB
Th M k t Di t i T iTh M k t Di t i T ihttp://bit.ly/KKDVh
Information f f i d
The Morakot Disaster in TaiwanThe Morakot Disaster in Taiwan
from friends, family and neighbors are particularparticular important for the public to fulfillfulfill information seeking requirements Within 5 minutes, one Taiwanese Within 5 minutes, one Taiwanese Plurker’sPlurker’s message message
di t b ild ll b ti di t h d bdi t b ild ll b ti di t h d brequirements, and to make decisions, because this kind of information provides local context, rapid updates as well as
regarding to build a collaborative disaster map had been regarding to build a collaborative disaster map had been spread to more than 6,500 people.spread to more than 6,500 people.
p , p psafety & welfare checking of close relationships.
Palen, L., S. Vieweg, et al. (2009). "Crisis in a Networked World."Social Science Computer Review 2009: 1-14.Huang & Chuang, 2010
48
An Event Ontology
lt ti i tlt ti i talternatives exist alternatives exist choices are yourschoices are yourschoices are yourschoices are yours
Communication ontology: http://sigmakee.cvs.sourceforge.net/viewvc/sigmakee/KBs/Communications.kif
S ti ll I t li k d O li C iti
Tag ontology: http://www.holygoat.co.uk/owl/redwood/0.1/tags/
1. General Event Relation (GER) is a conceptual level suggestion to serve as a common contextlevel suggestion to serve as a common context for disaster management communities to connect and utilize rich Web resources through the Event Conceptthe Event Concept.
2. While interest groups would like to develop their d i t l ld l lik town domain ontology, we would also like to
suggest some ontology guidelines when users need to build their own framework for specific needs. (see Appendix)
3. The main principles for linking crisis-disaster p p ginformation to web of data are three: Meaning, Relation, and Simplicity.
Huang & Chuang, 201051
An Event Ontology
http://guava.iis.sinica.edu.tw/ger/
M iM iM iM i R l tiR l tiR l tiR l ti Si li itSi li itSi li itSi li itMeaningMeaningMeaningMeaning RelationRelationRelationRelation SimplicitySimplicitySimplicitySimplicity
Huang & Chuang, 201052
An Event Ontology
ThanksThanksThanks Thanks becauseOf
YOUYOUYOUYOUHuang & Chuang, 2010
An Event Ontology
APPENDIX
Huang & Chuang, 201054
An Event Ontology
Huang & Chuang, 2010 Ontology for the Twenty First Century: An Introduction with Recommendations, 2006, Andrew D. Spear, Saarbrücken, Germany.
