Colloquia

Scientific Communication and the Semantic Web

Tim Clark
MIND Center for Interdisciplinary Informatics
(MassGeneral Institute for Neurodegenerative Disease)

October 18, 2007
JEC 3117 - 4:00 p.m. to 5:00 p.m.
Refreshments at 3:30 p.m.

Abstract:

Philosophers of science have described knowledge as "justified true belief". In biomedicine the contextual surround provides the "justification for true belief of a scientific statement, including the ability to replicate or extend results, and to place the statement within a larger body of theory. This context is composed of, among other things, training, experiment, and scientific discourse. Because today the crucial "discourse" part of this context must be gleaned from the text and is not machine-interpretable, it must be laboriously re-synthesized by the researchers in the process of scientific communication. Yet the content itself is almost all available on the web. In biomedicine, the contextual "surround" for this web-based content includes authorship, provenance, evidence, and discourse relationships. Discourse relationships include published hypotheses, open research questions, established theory, challenged theory, well-supported data linked to theory, and theory in process of formation. The SWAN project is an attempt to bind scientific communications content within a machine-interpretable formal semantic context for high bandwidth multi-faceted knowledge exchange. This formal semantic context is deliberately established in a way which does not presume the outcome of debates and theoretical struggles in any field. And it is intended to be established by the researchers themselves, as a side-effect of organizing their own knowledge resources. We intend to exchange and interlink knowledge and evidentiary support in digital communities, supporting a diversity of interpretations. Our current community template is the Alzheimer Disease Research community, as organized in Alzforum, a ten-year-old scientific web community with very strong acceptance and credibility in its field. So far the SWAN project has catalogued over 100 core and variant hypotheses in Alzheimer Disease, and is in process of deeply annotating 20-25 of them, as a "bootstrapping" dataset for researchers. This deep annotation includes argumentation structure, evidentiary support, argumentation relationships with other hypotheses in the field, and links to other biomedical ontologies. SWAN uses Semantic Web technology, including an OWL ontology, which we call an "ontology of discourse". All its data are stored in a triple store and accessed via SAPRQL queries. We will discuss the SWAN program and perspective, and demonstrate the SWAN Hypothesis Workbench and the SWAN browser in pre-beta versions.

Bio:

Tim Clark is a researcher in neuroinformatics, and Principal Investigator of the Semantic Web Applications in Neuromedicine project (SWAN). He is currently Senior Advisor and Core Member of IIC (Harvards Initiative in Innovative Computing), where he served as Founding Research Director in 2006-2007. He is appointed as an Instructor in Neurology at Harvard Medical School. Before coming to Harvard, Tim was Vice President of Informatics at Millennium Pharmaceuticals, where his team built one of the first integrated bio- and chemi-informatics software platforms in the pharmaceutical industry. Tim is a founding Editorial Board member of the journal Briefings in Bioinformatics; Director of Informatics at the MassGeneral Institute for Neurodegenerative Disease; and a co-Investigator at the Massachusetts Alzheimer Disease Research Center and the MGH/MIT Morris K. Udall Center of Excellence in Parkinsons Disease Research. He holds an M.S. in Computer Science from Johns Hopkins.

Hosted by: Jim Hendler (x4401)
Administrative support: Shannon Carrothers (x6354)

Last updated: October 1, 2007