> At 17:48 +0200 30/03/04, Nic Cottrell wrote: > >Has there been a lot of research into the possibility of using > >user-interactivity to improve the quality of MT? I was thinking of > >something like a spell checker but to assist in correct sense > >disambiguation etc. > >Any links or papers would be much appreciated.
Quoting Christian Boitet <[EMAIL PROTECTED]>: > A *main* point is that attempts to systematically ask questions while > the system is processing fail because users become "slaves of the > machine" and don't like it! Another is that users should not be > supposed to be specialists in grammar, nor to know the system, or the > target language(s). The following course was intentionally and specifically created ALLEN, Jeffrey. 1995. Grammar review course for technical writing. Peoria, Illinois: Technical Information Department, Caterpillar Inc. September - October 1995. (6-hour course) (Not publicly available) so that the following manual and training course would be more easily understandable for the users: JOHNSON, Paula and Jeffrey ALLEN. 1995. Caterpillar Technical English controlled language reference manual and Training course. Peoria, Illinois: Technical Information Department, Caterpillar Inc. September 1995. (Not publicly available) Human-Machine interactive identification of several kinds of attachments was a significant part of the training and mentoring process, up until automated attachment processing was enabled based on domain and semantic-based analyses. Several other articles on the topic: Baker, K., A. Franz, P. Jordan, T. Mitamura, and E. Nyberg (1994). ``Coping With Ambiguity in a Large-Scale Machine Translation System,'' Proceedings of COLING-94. (should be available on-line at: http://www.lti.cs.cmu.edu/) Mitamura, T., E. Nyberg, E. Torrejon, and B. Igo (1999). "Multiple Strategies for Automatic Disambiguation in Technical Translation". In proceedings of the 8th International Conference on Theoretical and Methodological Issues in Machine Translation (TMI-99) (should be available on-line at: http://www.lti.cs.cmu.edu/) Kamprath, Christine, Adolphson, Eric, Mitamura, Teruko and Eric Nyberg. 1998. Controlled Language Multilingual Document Production: Experience with Caterpillar Technical English. In CLAW98 proceedings, pp. 51-61. Available at: http://www.lti.cs.cmu.edu/Research/Kant/PDF/claw98ck.pdf (see http://www.controlled-language.org) Mitamura, Teruko. 1999. Controlled Language for Multilingual Machine Translation. Paper presented at MT Summit VII, Singapore, 13-17 September 1999. Available at: http://www.lti.cs.cmu.edu/Research/Kant/PDF/MTSummit99.pdf Mitamura, Teruko and Eric Nyberg. 1995. Controlled English for Knowledge-Based MT: Experience with the KANT System. Paper presented at 6th International Conference on Theoretical and Methodological Issues (TMI) in Machine Translation. Leuven, Belgium, 5-7 July 1995. Available at: http://www.lti.cs.cmu.edu/Research/Kant Nyberg, Eric, Kamprath, Christine and Teruko Mitamura. 1998. The KANT Translation System: from R&D to Large-Scale Deployment. In LISA Newsletter 2, no. 1 (March 1998). Available at: http://www.lti.cs.cmu.edu/Research/Kant/PDF/lisanews.pdf See several of Ralf Brown's papers on disambiguation and user interaction (all available at: http://www-2.cs.cmu.edu/~ralf/papers.html): Ralf D. Brown, "Augmentation", Machine Translation, 1989, vol 4 #2, pp. 129- 147. Ralf D. Brown, "Augmentation" in K. Goodman, ed. KBMT-89 Project Report. Center for Machine Translation, Carnegie Mellon University. 1989. Ralf D. Brown and Sergei Nirenburg, "Human-Computer Interaction for Semantic Disambiguation". In Proceedings of the Thirteenth International Conference on Computational Linguistics (COLING'90), vol 3, pp. 42-47. Helsinki, Finland Ralf D. Brown. "Automatic and Interactive Augmentation". In K. Goodman and S. Nirenburg (ed), The KBMT Project: A Case Study in Knowledge-Based Machine Translation. Morgan Kaufmann Publishers, 1991. ISBN 1-55860-129-5. Ralf D. Brown, "Improving Embedded Machine Translation with User Interaction", In Proceedings of the 1998 AMTA Workshop on Embedded Machine Translation, Langhorne, Pennsylvania, 28 October 1998. Chris Hogan and I also created a post-processing module to capture the results of MT Postediting decisions and use them for further MT and MT Postediting efforts. ALLEN, Jeffrey and Christopher HOGAN. 2000. Toward the development of a post-editing module for Machine Translation raw output: a new productivity tool for processing controlled language. In proceedings of the Third International Controlled Language Applications Workshop (CLAW2000), Seattle, Washington, 29-30 April 2000. (see http://www.geocities.com/mtpostediting/) An example of usability testing and changing a system accordingly for users is documented in: ALLEN, Jeffrey and Christopher HOGAN. 1998. Expanding lexical coverage of parallel corpora for the Example-Based Machine Translation approach. In Proceedings of the First International Language Resources and Evaluation Conference (LREC98), 28-30 May 1998, Granada, Spain. Vol. 2, pp. 747-754. (see http://www.geocities.com/mtpostediting/) > Questions should be asked > at some intermediate points, at the discretion of the users (e.g., > after all-path analysis) > in the source language > with straightforward questions (no trees, no notions of "PP-attachments�" > AND they should be asked to the proper persons. At IBM-Japan, the > technical writers agreed to answer lexical questions by JETS (J->E), > but not questions about dependencies between words. They felt it was > not their business -- and maybe the interface, although very nice, > was too "linguistic" for them. The key question is "how much is the customer willing to invest upfront and in ongoing maintenance for a system to automatically do these things for the users versus ordering a more basic version that requires more prerequisite linguistic knowledge on the part of the users?" You simply get what you pay for. And when the software and special features do exist, it is incredible how many people use the free or very inexpensive tools to try to attain the same quality output. I now never advise anyone to use free online MT portals to conduct any amount of MT postediting beyond a few paragraphs since there are several very good MT software programs which provide MT postediting environments, and which I have reviewed extensively (see http://www.geocities.com/jeffallenpubs/) . Yet it seems that there are people who use free or very inexpensive systems to avoid spending 300-500 US dollars/Euros on a professional or expert version of MT software that would really make their life easier. Usability testing is unfortunately one key area that often gets neglected in software development through implementation cycles. I am currently setting up training courses, train-the-trainer courses, and usability testing for a software product that has been deployed for the past 2 years in a different field. From this training and specific testing cycle, we are making a lot of usability-focused change requests to the development team. Usability testing should be a key component to both Controlled Language and MT products, whether they be commercial, academic or government developed. Regards, Jeff Jeff Allen [EMAIL PROTECTED] http://www.geocities.com/jeffallenpubs/ _______________________________________________ MT-List mailing list [EMAIL PROTECTED] http://www.computing.dcu.ie/mailman/listinfo/mt-list
