A Generic Software Framework for building Hybrid Ontology-Backed Models for Driving Applications

A Generic Software Frameworkfor building

Hybrid Ontology-Backed Modelsfor

Driving Applications

Colin Puleston, James Cunningham, Alan Rector

Bio-Health Informatics Group, University of Manchester

Hybrid-Model Demo

Demo Shows…

• Object Model (Java) provides central structure

• OWL Ontology provides detailed domain knowledge

• Dynamic model updates:– Orchestrated by Object Model– Based on interaction with ontology +

classifier

Object-Model/Ontology Interaction

EKS Access System

Object Model accesses EKS and associated inference mechanisms via mediating interface

…has no knowledge of any EKS format

…and no knowledge of any inference mechanisms

FaCT++

OWL-Access Protege-Ac.

Domain Knowledge Sources(ontologies, databases, etc.)

Ontological knowledge comes from set of External Knowledge Sources (EKS)…

Object Model

EKS Access System

Object Model has no prior knowledge of any EKS contents…

…other than an expectation that a small number of ‘key-entities’ will be represented somewhere

…and that mappings will be provided via a configuration file

FaCT++

OWL-Access Protege-Ac.

Domain Knowledge Sources(ontologies, databases, etc.)

Ontological knowledge comes from set of External Knowledge Sources (EKS)…

Object Model

EKS Access System

Object Model

Get concept-hierarchy

for<problem-

type-identifier>

Create instance

of C1 C1a

Update field-set +

constraints for

Cancer

Cancer selectedfield-value updated

C1

C1

Concept-Instance state update (OWL Implementation)…

Concept-Instance

Field Root Value

Cancer

locus Locus Breast

Instance-of

stage BC+s Stage2

subSt. BC2+ss BC2+ss

Concept-Instance

Field Root Value

Cancer

locus Locus Breast

Instance-of

stage BC+s Stage2

Cancerand locus some Breastand stage some Stage2

Stage2BreastCancerEquivalentTo: BreastCancer and stage some Stage2SubClassOf: subStage some (SubStageA or SubStageB…)

To OWL Description

Classify

Update(via ‘sanctioning’ mechanism)

Note

• EKS Access System allows ‘compositional’ concepts such as BreastCancer and Stage2BreastCancer to be hidden from Object Model

• …Hence: No need for classification during execution of queries constructed via Object Model

• …Hence: Relatively efficient (SPARQL-based) querying of large numbers of individual records

Why Hybrid Models?

type locus

Problem-Glimpse

Cancer Breast

subStagestage

Stage2 SubStageA

For simple GLIMPSE model of time…

...simple one-to-one correspondence between Object Model and EKS Access System

Concept-Instance

Field Root Value

Cancer

locus Locus Breast

Instance-of

stage BC+s Stage2

subSt. BC2+ss BC2+ss

stage

type locus

snapshot snapshot snapshot

stage subStagestage

Problem-History

Problem-Snapshot

Problem-Snapshot

Problem-Snapshot

Stage1 Stage1 Stage3 SubStageA

Cancer Breast

But… full SNAP/SPAN model of time involves non-ontologically-specifiable interaction…

min max

stage

type locus stage

snapshot snapshot snapshot

stage subStagestage

Problem-History

Problem-Snapshot

Temporal-Abstractions

Problem-Snapshot

Problem-Snapshot

Stage1 Stage3

Stage1 Stage1 Stage3 SubStageA

Cancer Breast

start end

Stage1 Stage3

Including interaction with a Temporal Abstraction System…

stage

type locus

snapshot

Problem-Snapshot

Stage1

Cancer Breast

Concept-Instance

Field Root Value

Cancer

locus Locus Breast

Instance-of

stage BC+s Stage1

Hence, more complex mappings between Object Model and EKS Access System…

Problem-History

type locus

snapshot

stage

Problem-Snapshot

Stage1

Cancer Breast

Concept-Instance

Field Root Value

Cancer

locus Locus Breast

Instance-of

stage BC+s Stage1

Hence, more complex mappings between Object Model and EKS Access System…

Problem-History

type locus

Cancer Breast

Concept-Instance

Field Root Value

Cancer

locus Locus Breast

Instance-of

stage BC+s Stage3

subSt. BC2+ss SSA

snapshot

subStagestage

Problem-Snapshot

Stage3 SubStageA

Hence, more complex mappings between Object Model and EKS Access System…

Problem-History

min max

type locus stage

Temporal-Abstractions

Stage1 Stage3

Cancer Breast

start end

Stage1 Stage3

Concept-Instance

Field Root Value

Cancer

locus Locus Breast

Instance-of

stage BC+s ???

Temporal-Abstractor-Set

+ Field Type

stage ORDINAL

…also involving Temporal Abstraction System

Hence, more complex mappings between Object Model and EKS Access System…

Problem-History

Hybrid Models Summary

• Non-ontologically specifiable interaction• Hence, hybrid models of some type

required• Advantages of Object-Model based

solution:– Relatively simple– Flexible– Coherent– Domain-specific API

Higher-Level Model Architecture

Higher-level interaction orchestrated by Interaction Model…

Object Model

Static Model(SNAP/SPAN)

Interaction Model

EKSAccessSystem

Temporal Abstraction

System

Additional Processing Mechanisms distinct from Interaction Model - do not contribute to shape of model itself

Object Model

Static Model(SNAP/SPAN)

Additional Processing

Mechanisms

Interaction Model

EKSAccessSystem

Temporal Abstraction

System

Driving Applications

•Domain-specific API

•Interaction Model intrinsic to domain classes (extending generic patterns where applicable)

Basic Source Representation provided by collection of domain-specific Java classes…

A

B

E

C

D

V

ZX

Y

W

7

4.3

28/1/97

Field on objectAObject of (domain-specific) class A

Relationship represented by object of class Link

Automatic translation into domain-neutral Network Representation…

•Domain-neutral semantic network representation

•Interaction Model translated into a set of ‘listeners’

•All translation specification is by individual classes

A

B

E

C

D

V

ZX

Y

W

7

4.3

28/1/97

A Object of class Node - representing instance of concept ‘A’

Relationship represented by object of class Link

B

C

D

E

Z

X

Y

?q

?p

OR

F

OR

< (?p + ?q) / 2

> ?pA

NOT

Query formulation via extended version of Network Representation…

Extensions for…

•Insertion of logical operators

•Specification of data-value constraints (including temporal constraints)

•Specification of temporal-abstractions over dynamically defined time-periods (not shown)

Software

Patient Chronicle Model

EKS Access System

Core Model-Builder

Chronicle Model-Builder TemporalAbstraction

System

Core Model-Builder provides generic EKS-interaction mechanisms, skeleton interaction-patterns, etc.Chronicle Model-Builder provides core SNAP/SPAN representation, interaction with Temporal Abstraction System, chronicle-specific interaction-patterns, etc.

OWL-Access Protege-Ac.

Model

Storage Manager

RDF Store

Data-Storage/Query System (Domain-Neutral)

GUIs (Domain-Neutral)

Record Browser

Model Browser

Query Formulator

Patient Record Chronicliser

EHR Store

Data-Creation Tools (Domain-Specific)

Patient Chronicle Simulator

QueryEngine

Temporal AbstractionExtension

Further demos available on request– please ask!

Software soon available on-line(Technical Supplement will provide details)