18
Implementing a Role Based Mutual Assistance Community with Semantic Service Description and Matching Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

  • Upload
    leona

  • View
    30

  • Download
    0

Embed Size (px)

DESCRIPTION

Implementing a Role Based Mutual Assistance Community with Semantic Service Description and Matching. Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS. Forecast. - PowerPoint PPT Presentation

Citation preview

Page 1: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

Implementing a Role Based Mutual Assistance Community with Semantic Service Description and Matching

Hong Sun, AGFA HealthcareVincenzo De Florio, Chris Blondia, University of Antwerp, PATS

Page 2: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

ForecastForecast

MEDES 2013 30 October, Luxembourg 2

The proportion of elderly people keeps increasing and poses challenges in healthcare in providing cost effective services.

Our answer: constructing a role-based mutual assistance community (MAC) to fully utilize the community resourcesand improve the physical and psychological health of the assisted people.

Cornerstone of MAC: Service publication and matching

An architecture based on semantic service description and matching is presented in this paper.

Page 3: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

OutlineOutline

Role-based MAC: Introduction

Semantic service description and matching for MAC

Experiments and performance

Conclusions

MEDES 2013 30 October, Luxembourg 3

Page 4: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

Mutual Assistance CommunityMutual Assistance Community

MEDES 2013 30 October, Luxembourg 4

Page 5: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

Mutual Assistance CommunityMutual Assistance Community

MEDES 2013 30 October, Luxembourg 5

A community dweller is encouraged to act as service provider

A community dweller may play different roles

- so as to removes any predefined and artificial distinction between care-givers and care-takers

A community dweller is encouraged to participate group activities

Community resources are therefore better utilized

Page 6: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

Previous OWL-S Based Previous OWL-S Based ImplementationImplementation

MEDES 2013 30 October, Luxembourg 6

Setting MatchingRequirements

Page 7: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

Drawbacks of OWL-S Based Drawbacks of OWL-S Based ApproachApproach

MEDES 2013 30 October, Luxembourg 7

Strict and inflexible format of service expression and matchingThe majority of OWL-S matchmakers are based on service inputs/outputs impossible to match complex services.

Performance penalty in service search/comparisonOWL-S Matcher takes more than 6 seconds to load the profile for service publication, and it takes around 6 seconds to compare two services.

OWLS-MX takes more than 1 second when the service size is around 400.

Page 8: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

RDF Graph Based Service RDF Graph Based Service Description/MatchingDescription/Matching

MEDES 2013 30 October, Luxembourg 8

Flexible in service description

Easy in service managementservice publication, update, query, termination

Fast in service matching

Service description as RDF graph, which is

Service management/query is coordinated by Fuseki SPARQL server

Page 9: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

Basic OntologyBasic Ontology

MEDES 2013 30 October, Luxembourg 9

@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>.@prefix service: <http://www.pats.ua.ac.be/AALService#>.@prefix owl: <http://www.w3.org/2002/07/owl#>.

service:Activity a owl:Class.

service:ServiceCategory a owl:Class.

service:Fitness a owl:Class; rdfs:subClassOf

service:ServiceCategory.

service:Cycling a owl:Class; rdfs:subClassOf

service:Fitness.

Service:Walking a owl:Class;

rdfs:subClassOf service:Fitness.

Page 10: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

Role Based Service DescriptionRole Based Service Description

MEDES 2013 30 October, Luxembourg 10

service:playRole defines the role that a community dwellers plays.service:hasScope defines the scope (service activity) that a role is bond to.

service:playRole

service:hasScope

roleservice content

service content

community dweller

role

Page 11: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

InferenceInference

MEDES 2013 30 October, Luxembourg 11

{?C rdfs:subClassOf ?D. ?D rdfs:subClassOf ?E} => {?C rdfs:subClassOf ?E}.

#service category inference rule { ?service service:hasServiceCategory ?serviceCategory. ?serviceCategory rdfs:subClassOf ?serviceCategoryClass. ?serviceCategoryClass log:notEqualTo service:ServiceCategory. } => { ?service service:hasInferredServiceCategory ?serviceCategoryClass. }.

<…/Activity/101> service:hasInferredServiceCategory service:Fitness.

<…/Activity/101> service:hasServiceCategory service:Jogging.

It enables to match services which are syntactically different, but semantically close. It is executed by EYE Reasoning Engine.

infer

Ref: EYE engine: http://eulersharp.sourceforge.net/

Page 12: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

ArchitectureArchitecture

MEDES 2013 30 October, Luxembourg 12

SPARQL Endpoint

Rule Engine

Query Generator

Activity Publisher

Web Portal

Repository

Service Lifecycle Management

Link Open Data

Page 13: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

WorkflowWorkflow

MEDES 2013 30 October, Luxembourg 13

query existing service request

group activity

No

invokeprovide service

User

publish

request service?

as group activity

?

query existing group activity

exist?query existing available service

Yes

exist?exist?

exist?

Nopublish

No

select

bind

select

bindselect

bind

select

bind

publish No

Yes

Yes

Yes Yes

No

Page 14: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

Service QueryService Query

MEDES 2013 30 October, Luxembourg 14

prefix service: <http://www.pats.ua.ac.be/AALService#>prefix xsd: <http://www.w3.org/2001/XMLSchema#>

DESCRIBE ?creatorWHERE { ?service ?servicetype ?type. ?service service:hasCreator ?creator . ?creator service:playRole ?role. ?role ?serviceRole ?roleType. ?role service:hasScope ?service. ?service service:endTime ?endTime. FILTER(?endTime > "2013-05-01T17:00:00"^^xsd:dateTime) FILTER(?type = service:Fitness) FILTER(?roleType =service:GroupAcitivityParticipant)}

It is possible to filter on different properties; the service match process is therefore much more flexible compared with our previous implementation.

Page 15: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

Service Publication and Service Publication and TerminationTermination

MEDES 2013 30 October, Luxembourg 15

prefix service: <http://www.pats.ua.ac.be/AALService#>prefix xsd: <http://www.w3.org/2001/XMLSchema#>

INSERT DATA { #Service Description, as line 3– line 14 in Table 1.<http://www.pats.ua.ac.be/aal/Activity/10100000#this> a service:Activity....}

1 prefix service: <http://www.pats.ua.ac.be/AALService#>2 prefix xsd: <http://www.w3.org/2001/XMLSchema#>

3 DELETE { 4 ?service ?p ?o.5 ?o ?p2 ?o2.6 ?o2 ?p3 ?o3.7 } where {8 ?service service:endTime ?endtime.9 FILTER ( ?endtime < "2013-05-26T00:00:00"^^xsd:dateTime )10 ?service ?p ?o.11 OPTIONAL {?o ?p2 ?o2. OPTIONAL {?o2 ?p3 ?o3.}}12 }

Service Publication Service Termination

Page 16: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

PerformancePerformance

MEDES 2013 30 October, Luxembourg 16

target graph size

activities

triples

time

thread 1

thread 5

1000

12000

16.2ms

44.2ms

10000

120000

99.8ms

291.5ms

50000

600000

638.8ms

1.82s

100000

1200000

1.36s

4.14s

150000

1800000

2.17s

6.91s

200000

heap space

exceptionthread 10 81.1ms 616.3ms 3.99s 8.05s 13.95s

250000

2400000

18.75s

115.38s

250s

3000000

A set of sample activity graphs that defines a different number of activities are randomly generated.

Different threads of service queries are tested as well.

Page 17: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

……Every new solution breeds new Every new solution breeds new problems…problems… MAC basic assumption: every new service

needs to be matched with known services scalability

Upper bound on community size (heap space exception)

Current solution: breaking down the communities into sub-communities

Fractal organization

MEDES 2013 30 October, Luxembourg 17

Page 18: Hong Sun, AGFA Healthcare Vincenzo De Florio, Chris Blondia, University of Antwerp, PATS

ConclusionConclusion

Semantic service description & matching framework for MAC

MEDES 2013 30 October, Luxembourg 18

• Flexibility in service description.

• Ease in service management.

• Speed in service matching.

RDF graph based implementation outperforms our previous OWL-S based implementation in:

Future work: 1) semantic service composition2) automate the workflow3) fractal social organizations

http://www.youtube.com/watch?v=HKU7gOI57e0