Ontology Use for Semantic e-Science

Paper Title: 
Ontology Use for Semantic e-Science
Boyan Brodaric, Mark Gahegan
Ontologies are being widely used in online science activities, or e-Science, mostly in roles related to managing and integrating data resources and workflows. We suggest this use has focused on enabling e-science infrastructures to operate more efficiently, but has had less emphasis on scientific knowledge innovation. A greater focus on online innovation can be achieved through more explicit representation of scientific artifacts such as theories and models, and more online tools to enable scientists to directly generate and test such representations. This should lead to regular use of ontologies by scientists as part of their routine online activity.
Full PDF Version: 
Submission type: 
Responsible editor: 

Review 1 by Michel Dumontier:
This well written article presents a vision for how e-Science can be brought to its next level by not just using ontologies for sharing information and facilitating data integration, but rather to be used to construct and validate scientific hypotheses over highly structured descriptions of methods and their resulting data. In particular, the authors argue that provenance needs to extend past keeping track of where the data came from (source-based), but to really have a representation of the scientific method so as to have structured representations of their goals and to be able to evaluate them. Finally, they argue that systems that aim to capture scientific discourse need to be formalized so as to leverage the facts found in ontologies and databases.

All in all, i like the general messages being put forward in this paper, they are certainly on track and worth pursuing :-)

Review 2 by Manfred Hauswirth:
The paper argues for the use of ontologies in e-science in a more comprehensive way, i.e., beyond the current uses in relation to resource efficiency, data interoperability, etc. which I would see as "traditional" computer science problems. The authors propose to go beyond this purely technical use of ontologies as a means to express hypotheses, claims, proofs, etc., i.e., the elements of scientific discourse and knowledge generation, but well-supported by e-science environments which (semi-) automatically support the generation of machine readable and interpretable representations of scientific knowledge. This representation could then be used detect contradictions, new knowledge (by combining existing and inferring new knowledge), better recommending related knowledge to scientists, etc.

While I think such a system would be very useful, the authors do not provide a lot of hints how such a system may be achieved. A (rough) roadmap may help in this respect to demonstrate what issues need to be addressed to come up with a system as envisioned by the authors.

It may also be interesting to look into related work that has tried to apply ontologies for the purpose of e-science. This work has the potential to be translated into the authors' vision of future e-science. For example, the work on SALT [1] and Konnex [2], which establishes a infrastructure for the explicit representation of knowledge artifacts within scientific publications. In particular it does (i) finding claims in scientific publications, and (ii) building the argumentation discourse network (ADN) for each claim and providing support for browsing it, by making use of transclusion. On a first glance, this seems to be highly relevant for your vision.

[1] SALT - Semantically Annotated for Scientific Publications. In The Semantic Web: Research and Applications , Vol. 4519 (2007), pp. 518-532. by Tudor Groza, Siegfried Handschuh, Knud Möller, Stefan Decker

[2] KonneX-SALT: First Steps towards a Semantic Claim Federation Infrastructure. In the Proceedings of ESWC 2008. by Tudor Groza, Siegfried Handschuh, Knud Möller, Stefan Decker