semantic search thoughts

Thomas Tague has asked for input to help better define semantic search.

So, let’s deconstruct semantic search into it’s constituent components and talk a bit about how and whether semantic technologies might actually make it better.  The results of the dissection are here on the table….

   1. What kinds of questions can we ask? Can we embed logic in our questions? Do we expect inference in our results?
   2. How can we ask them – keywords, natural language and all that jazz.
   3. Generating the “right” result set for the query.
   4. Displaying the result set in the most effective manner
   5. Making money from doing all that

Here at CISTI we see Semantic Search as a key advanced technology to help the library support researchers, and in particular to help researchers get at the meaning locked up in our millions of locally-loaded journal articles.

We think of it as three pieces:

1. Domain-specific semantic extraction, in which the different types of scientific information are identified and indexed for search (e.g. chemical names, mathematical formulas, genes, etc.)
2. Straightforward search frontends, which use basic keyword mapping to provide an enhanced discovery interface (e.g. you search for water and you get any documents with the various possible representations for water)
3. Semantic search frontends (possibly also domain-specific) in which a researcher enters a search in terminology that's meaningful to them - for example drawing a chemical structure, or asking a complex question (e.g. what X are affected by Y when conditions A and B hold)

For all of these areas we see partnering as a key, as no single organisation is going to have the domain-specific expertise necessary, or the technology expertise necessary.

I'd be happy to hear other opinions on semantic search, particularly as it relates to science.
Should we separate it into several subareas?  Perhaps

• Semantic embedding (the document is already semantically enriched)
• semantic extraction (terms are extracted from a document that isn't semantically encoded)
• semantic indexing (constructing search indexes using semantic information)
• semantic search (search engines that use semantic information)

I see I don't even have a category for Semantic Web, so I guess it's time to make one.