Wednesday, May 16, 2012

Crazy about Semantics

Last March I wrote a post about Google’s plan for “a phased transition from its keyword-based search system into a technology that is more ‘semantic.’” Drawing upon my own study of semantics and my familiarity with the work of others that have been trying to crack this nut, I gave the post the title “Google’s Assault on Semantics.” The intent behind the title was the connotation that Google was more interested in the big stick of authority than in all those complexities that have made the study of semantics so problematic.

According to a CNET News article by Rafe Needleman that appeared this morning, that transition is now under way. Google has introduced a new approach to the display of search results, which they call “Knowledge Graph.” Those familiar with graph theory (and, perhaps, its current applications to social networks) are likely to be disappointed. Any sense that a graph has something to do with connections between entities has been displaced by a display of results based primarily on colocation. As to the “knowledge” part of the title, given the extent to which technology has robbed that word of a legacy of meaning that reaches back millennia, it would suffice to say that Google has just added another brick to the wall.

My March post concluded with the following jaundiced observation:
Thus, a truly “semantic” system will have to have a fair amount of knowledge about your personal psychology and probably the sociology of that corner of the world you inhabit.  Can this be done?  It certainly is a challenging research question, and Google probably has the resources to support an appropriate research program.  However, the more important question is:  Do you really want Google to have a model of your psychology and the sociology of your world?  If you think that advertising is already invasive, imagine what it would be like if Google could tap into your personal psychology and sociology!
This should provide a useful context for considering what Google currently thinks of semantics. In Needleman’s article those thoughts were voiced by Jack Manzel, Product Management Director of Search:
We do continue to work on how to make search semantic but talking about it brings out the crazy people.
My guess is that the slightest suggestion of anything that cannot be readily reduced to what can be expressed in a programming language is immediately classified as “crazy.” That would include both psychology and sociology. Given my concluding cautionary warning, it may be just as well that those willing to recognize these as potentially valuable areas of research are getting summarily dismissed as “crazy!”

1 comment:

Stephen Smoliar said...

It appears that Google did not want me to moderate a comment from jones (again); so I figure it is worth reproducing as a comment of my own:
> This should provide a useful context for considering what Google currently thinks of semantics

Or why Google launched Google+

Facebook is a parasite: if a facebook user links to a website, Facebook gets to know who linked to it and who clicks on it, but the website doesn't, because facebook is not part of the "public web."

Given that Facebook accounts for as much traffic as Google now, this would seem to represent a significant hurdle for google's semantic search plans