ECDL 2006 - DLSci06 - Provenance Explorer

Kwok Cheung and Jane Hunter
Provenance Explorer - A Tool for Viewing Provenance Trails and Constructing Scientific Publication Packages

* Modelling scientific discovery process
* Tools for constructing and publishing scientific models

Scientific Publishing - increasing demands to
* publish raw and derivative data
* document precise provencance
* share data models
* enble duplication and validation
* protect IP
* facilitate training / learning objects

Huge number of elements you should capture to enable reproduction of experiments

Huge Knowledge Management Challenges
* very large data set
* distributed data sets
* etc.

Need to capture the the provenance

* may display/share various levels of detail
* selective archiving

Lineage data/metadata

* workflow is prospective
* provenance is retrospective

Capture the Semantic Descriptions of components

FUSION project

http://metadata.net/sunago/fusion.htm

ontologies: ABC, MPEG-7, FUSION, OME

Image labelling

* Rules-by-example
  - then you can automatically generate semantics

Based on built-up semantics, you can generate automatic displays.

Hypothesis testing interface

Also, can specify new areas of interest e.g. by identifying sparse areas in the 3D display of results.
This can automatically generate parameters for new experiments.

Extended Harmony ABC model for experiments.

http://jodi.ecs.soton.ac.uk/Articles/v02/i02/Lagoze/

Result - extended ABC

Then: Modelling eScience Provenance

[various screenshots of the application]

Architecture

* JGraph
* Algernon inference engine
* Protege OWL
* Jena semantic web framework

Working on

* Scientific Model/Publication construction tools
* Search, Browse and Retrieval

[diagram of architecture]

Preservation of Composite Objects
* Use RDF/XML to package metadata
* Maintain preservation for both
  - composite objects
  - atomic objects

PANIC Architecture

PANIC (Preservation webservices Architecture for Newmedia and Interactive Collections)

http://metadata.net/panic/

AONS

* Automated Obsolescence Notification Service

* Collaboration between
  - UQ
  - NLA
  - ANU?