[developers] Paraphrase grammars and SEMI's

Francis Bond fcbond at gmail.com
Mon Jun 1 08:52:38 CEST 2009


Thank you everyone for your replies.   Consensus seems to have been
reached, I am summarizing here, and have also added it to the wiki at

The monolingual machine translation should use the external MRS
exposed by the Semantic Interface (SEM-I), not the internal MRS, for
the following reasons:

   1.   the SEM-I is the interface to outside resources, such as
WordNet, which may also be used in generation
   2.   the internal MRS is subject to restructuring as the quest for
ever better linguistic analysis proceeds
   3.   the mtr.tdl used in MT refers to the external MRS

The SEM-I should, therefore, expose enough information for the
paraphrase engine, including some hierarchies among types, this must
be done manually at present.


Francis Bond <http://www2.nict.go.jp/x/x161/en/member/bond/>
NICT Language Infrastructure Group

More information about the developers mailing list