We have had some discussion about schema.org as a kind of minimal
ontology sponsored by Google, Bing, and Yahoo! But there are much,
much larger ontology-like things that Google, Facebook, and others
have been developing. (01)
Last May, Google announced their "Knowledge Graph", aspects of which
have been appearing on the right-hand sides of their search results.
Their catch phrase is "things, not strings": (02)
http://googleblog.blogspot.co.uk/2012/05/introducing-knowledge-graph-things-not.html (03)
In December, Google announced that their graph has grown from 3.5
billion "facts" to 18 billion facts in seven months. But the number
of "things" mentioned in those facts grew at a slower pace --
from 500 million "objects" to 570 million "entities": (04)
http://news.cnet.com/8301-1023_3-57557055-93/googles-knowledge-graph-tripled-in-size-in-seven-months/ (05)
Last week, Mark Zuckerberg introduced Facebook's Graph Search: (06)
http://news.cnet.com/8301-1023_3-57564611-93/what-should-google-do-about-facebook-graph-search/ (07)
From that article,
> The service incorporates various filters such as "place type,"
> "liked by," and "visited by friends" to make locating things faster.
> You can refine search queries with more advanced filters to get better
>answers.
>
> One example demonstrated was a very specific search for "Friends of
> my friends who are single male San Francisco, Calif." That refined query
> returned a select group of people who fit the criteria. (08)
In logic, we would call those simple "filters" relations. The advanced
"filters" are controlled English that could be translated to logical
expressions. (09)
The Facebook graph is a giant graph DB, and their version of controlled
English is far more readable and writable than SQL or SPARQL. (010)
Google's collection of documents is much, much larger and far less
structured than Facebook's graph. It would be much harder to support
DB queries of the kind that Facebook processes. (011)
However, Google maps does support some NL-like queries. For example,
type "from grand central to penn station" to Google. I was surprised
that Google's Knowledge Graph did not post any side information
about "grand central" or "penn station" as named entities. (012)
Instead, the first hit was a USA Today site that gave directions
on how to go by subway or by taxi: (013)
http://traveltips.usatoday.com/grand-central-station-penn-station-13435.html (014)
Then I clicked on Google maps, which gave directions on how to go
by walking or by driving. However, I can't imagine anybody who
asked that question who would have a car at GCT. (015)
Then I tried Bing.com, for which the first hit was by ehow: (016)
http://www.ehow.com/how_6163210_grand-penn-station-new-york.html (017)
This page listed four options: walking, taxi, subway, and bus. (018)
Then I tried Bing maps, but they were clueless. They asked
> By "grand central", did you mean:
>
> 1 Grand Central, Los Angeles Co., CA
> 2 Grand Central, Sangamon Co., IL (019)
When I clicked on #1, Bing told me that it would take
39 hours and 39 minutes to drive to Penn Station. (020)
John (021)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (022)
|