There are many constrained language situations where machine language
understanding would be extremely valuable, and possibly not too hard to do.
In such situations there is a good informal semantics. But one needs a
standardized precise semantics. For example, in the Language of Data
example I mentioned earlier, there were two different semantic contexts for
petroleum policy, one from the Department of Commerce, and one from the
Department of Energy, both fairly precise. The problems arose where data
from the two different sources was aggregated without taking into account
the different meanings. (01)
Consider the issue of policy for managing petroleum reserves. We are
currently aware of changes in the price of petroleum, as the price of
gasoline affects our daily life. Most people know that there are different
classifications of petroleum that sell at different prices, but probably do
not know much more detail. Government policy and planning for managing
petroleum reserves needs to be based on accurate information regarding
quantity of reserves, imports, and consumption. A precise characterization
of exactly what is petroleum is a prerequisite for determination of the
country’s reserves and import needs. Petroleum is made up of hydrocarbons;
however, which hydrocarbons are considered to be petroleum has a tremendous
impact on the meaning of statements regarding the magnitude of oil reserves.
Tar sands have only recently been considered to be part of the world's oil
reserves, as higher oil prices and new technology enable them to be
profitably extracted and upgraded to usable products. The inclusion of tar
sands makes a significant change to stated petroleum reserves. Accounting
for reserves of other substances, such as coal, which can be converted to
petroleum products changes our assessment even more. Some substances
require more or less effort and cost to transform into usable products than
others. Transformation of different substances has different impacts on the
environment. How do we know or insure that experts use the same terms or
use them in the same way? The consequences of misinterpretation can be
significant. Analysis of petroleum reserves requires precise and careful use
of terminology if it is to be of value in determining policy. One often
feels that reports on such topics are suspect and have been “cooked” to
support some conclusion that has already been made independently of the
evidence presented. The ordinary person seldom has the resources to get to
the bottom of the issue. Common sense is insufficient. A standard that can
be taught, applied, and checked by independent parties is needed, as are
information systems which can represent this data together with its
semantics, and enable data to be aggregated from different sources without
losing its meaning. There are lots of other similar examples. (02)
In my opinion such constrained language situations are a fruitful area of
research and development. It seems likely that a logic-based representation
would work for each agency. A translator or unification of the theories of
these individual representations would be needed. It is unclear to me
whether much of the forum discussion addresses these issues. (03)
Henson Graves (04)
_________________________________________________________________
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 (05)
|