How Google Does Synonyms

Fascinating post over at the Official Google Blog about how their search engine deals with synonyms. One of the more interesting points is about how they deal with "bad" synonyms:

An example of a bad in the search [dell system speaker driver precision 360], where Google thinks "pc" is a synonym for precision. Note that you can still see that on Google today, because while we know it's a bad synonym, we don't typically fix bad synonyms by hand. Instead, we try to discover general improvements to our algorithms to fix the problems. We hope it will be fixed automatically in some future changes.

I infer from this that Google doesn't typically fix any bad search results by hand. They would much rather fix the underlying algorithm than hard code a bunch of corner cases. In fact they'd so much rather do that, that the best they can say in this situation is that they "hope it will be fixed" someday soon!

At first glance you might think that Google makes this choice because it's right architecturally--because it's cleaner and easier maintain. But I think the bigger reason is that Google cannot be seen to be tipping its hand.

If Google were to manually fix the PC/precision synonym for instance, people could potentially argue that they are unfairly favoring Dell's search results over any others. But as long as they can always fall back on the "it's the algorithm" defense, Google's results can be seen as more impartial and therefore more relevant.

And, indeed, this is the fundamental asset that Google owns: your confidence in the impartiality and relevance of its search results.

Read Official Google Blog: Helping computers understand language.