fass@fornax.UUCP (Dan Fass) (10/11/90)
I am looking for post-1987 information about the menu-based approach
to natural language understanding (NLU) developed by Harry Tennant
(Tennant, 1987; Tennant et al, 1983).
Does anyone have the references for recent papers about the approach?
What are the advantages and disadvantages of the approach?
Has anyone had experience with the NLMenu System developed by Texas
Instruments?
Does anyone have Harry's e-mail address?
Please send mail to me and I will post a summary of responses to the net
(if there are any).
** -- Brief Description of the Menu-Based Approach to NLU -- **
The menu-based approach to NLU combines the expressive power of natural
language with the ease of use of menus.
In the pure version of the approach, sentences are built through menu
selection:
``The user is presented with a set of menus on the upper half
of a high resolution bit map display. He can choose the words
and phrases that make up his query with a mouse. As the user
chooses items, they are inserted into a window on the lower
half of the screen so that he can see the sentence he is
constructing. As a sentence is constructed, the active menus
and items in them change to reflect only the legal choices,
given the portion of the sentence that has already been input.
At any point in the construction of a natural language sentence,
only those words or phrases that could legally come next will be
displayed for the user to select'' (Tennant et al, 1983, p. 152).
The approach offers some attractive features as a natural language
interface (NLI):
o easy to use;
o cuts down on spelling and typographical errors;
o minimizes typewriter key/mouse operations,
o helps overcome user difficulties in starting a query;
o keeps users within the linguistic and conceptual coverage of the NLI;
o reveals to users the full coverage of the NLI through the menus.
Tennant et al (1983, p. 157) mention that the NLMenu System is implemented
in Lisp and that a second implementation will be available as a software
package that ``will interface either locally to RSI's Oracle relational
DBMS which uses SQL 3.0 as the query language or to remote computers
running DBMS's that use SQL 3.0 as their query language.''
** -- References -- **
Tennant, H. R., K. M. Ross, R. M. Saenz, C. W. Thompson, & J. R. Miller.
Menu-Based Natural Language Understanding.
In: Proceedings of the 21st Annual Meeting of the Association for
Computational Linguistics, MIT, Cambridge, MA, pp. 151-158, 1983.
Tennant, H. R.
Menu-Based Natural Language.
In: S. C. Shapiro (Ed.) Encyclopedia of Artificial Intelligence.
New York, NY: John Wiley & Sons, pp. 594-597, 1987.
______________
Dan Fass
fass@cs.sfu.cafass@fornax.UUCP (Dan Fass) (12/05/90)
Annette Myjak (arm@sps.com) wrote that > i used to work in the same lab at texas instruments as harry did. unless > he's relocated from the lab (doubtful, unless he's left ti) or has a really > different user id, his netmail address should be: tennant@csc.ti.com Graeme Hirst (gh@cs.rochester.edu) told me that There is some controversy (and bad feelings) about who actually invented the idea. I think Jim Hendler thinks it was his. He's now at U Maryland, hendler@cs.umd.edu. Jim Hendler (hendler@mimsy.umd.edu), one of the developers of the NLMenu system at Texas Instruments, sent a message suggesting I contact Brian Phillips, another of the developers, but I have been unable to locate him. He apparently works for Tektronix. Jim Tyhurst (jimt@context.mentor.com) mentions a paper that discusses some of the problems associated with the NLMenu system: Tyhurst, James J., and Kerry L. Glover. 1988. A menu-based interface for expert system rules. In Proceedings of the 2nd Annual Expert Systems Conference and Exposition (April 12-14, 1988). Detroit: Engineering Society of Detroit. pp. 203-210. Thomas Grossi (grossi@capsogeti.fr) told me that he has developed a front end quite similar to Tennant's. The front end is part of Esprit-I project "Esteam" and produces Functional Descriptions for a dialogue module (Functional Descriptions are a frame-based linguistic representation used in unification grammars). > The system is quite similar to Tennant's, in that at each step the user > chooses the next part of the sentence via a menu, except that I have > introduced "intermediary" menus that contain non-terminals. The sequence > of menus presented reflects the structure of the underlying functional > descriptions, in fact. A big plus in my system w.r.t Tennant's is that I > was able to use the same rules to derive a surface sentence from a > functional description; thus I have obtained "menu-based natural language > understanding and generation". Steve Cousins (sbc@informatics.WUstl.EDU) wrote > I did a master's thesis in 1987 incorporating some of the ideas of NL Menus. > It was entitled "Grammar-Based Techniques for Interface Design", and was > done at Washington University in St. Louis. A very consise (some would say > cryptic) version of the work appeared in the SIGART Newsletter in October > '87. My work describes a Prolog program which does much of what NL Menus > does in about 1/2 page of Prolog code (which modifies the DCG mechanism of > prolog). > You should also probably look at CW Thompson's PhD dissertation, which is > all about NL Menus. It's got much more detail than the Conference > proceedings you cited in your note. > Here are the references: (bibtex form) @phdthesis{thompson, author = "Craig Warren Thompson", title = "Using Menu-based Natural Language Understanding to Avoid Problems Associated with Traditional Natural Language Interfaces to Databases", school = "The University of Texas at Austin", year = 1984 } @article{cousins, author = "Steve Cousins", year = "October, 1987", journal = "SIGART Newsletter", title = "Automatic Menu Generation" } @misc{tennant, author = "Harry R. Tennant and Kenneth M. Ross and Craig W. Thomson", booktitle = "Proc. CHI '83 Conference on Human Factors in ComputingSystems", year = "1983", pages = "154-160", title = "Usable Natural Language Interfaces Through Menu-Based Natural Language Understanding" } @manual{ti-nlmenus, year = "1985", organization = "Texas Instruments", title = "Explorer Natural Language Menu System User's Guide" } Thank you to all those who sent me messages. I would welcome further information from people about menu-based NLU. If I receive more replies, I will summarize to the net again. _______________ Dan Fass fass@cs.sfu.ca
fass@fornax.UUCP (Dan Fass) (12/05/90)
Rene Bach (rene@tech.ascom.ch) wrote > I would check with GENSYM. G2 uses such an interface to write rules. > They are located in Cambridge, Mass. _______________ Dan Fass fass@cs.sfu.ca