Research methods and system designs in machine translation: a ten-year review, 1984-1994
What it is about
A review of MT research over ten years, organised by the five eras of the field. It describes the state in 1984, when SYSTRAN was in use at several large organisations and the Meteo weather system had been running since 1976. Until the end of the 1980s the rule-based approaches dominated, with transfer and interlingua designs such as Ariane and Eurotra. From 1989 the corpus-based approaches took over the research agenda: statistical methods from IBM's Candide project, and example-based methods. The review says the largest surprise was that a statistical system gave acceptable output, with almost half of the phrases matching the corpus or expressing the same sense. It predicts that future systems will be hybrids of rule-based, statistical and example-based parts, with corpora refining the lexical and semantic choices of the linguistic stages.
Ideas
- I1. Hybrid by layer: keep structure from the rule stack and take lexical choice from counts, as the review predicts. In the translator, measure which layer carries most of the residual bits per class, and give the count-based part only the classes where it wins.
- I2. Phrase match rate: report the share of output phrases found verbatim in the training pairs, per author, as Candide reported its phrase match. It shows how much of the output is copied from data rather than produced by the standard.
- I3. Same sense, different words: the review's second measure is a sense match rather than an exact match. Add a sense-level score to the chrF output, based on the meta concepts, so that a correct meaning in other words is counted.
For Mova
The review supports the layered design and the count-based refinement of lexical choice, which the standard encoder already uses. The phrase-match rate is a cheap measure to add per author. The sense-level score is the closest thing to the meta-world's own measure and would separate meaning errors from wording errors. None of this is a result for Mova yet.
Cite
@inproceedings{hutchins1994research,
title = {{Research methods and system designs in machine translation: a ten-year review, 1984-1994}},
author = {Hutchins, John},
booktitle = {Proceedings of the Second International Conference on Machine Translation: Ten years on},
year = {1994},
url = {https://aclanthology.org/1994.bcs-1.4/}
}References
- The sources for this paper are mainly the proceedings of the biennial MT Summit conferences (Hakone 1987, Munich 1989, Washington 1991, Kobe 1993) and of the international conferences on Theoretical and Methodological Issues in Machine Translation (TMI) at Austin 1990, Montreal 1992, and Kyoto 1993. For details of systems before 1990 see references in Hutchins 1986 and 1988. find
- AAAI (1993): Building lexicons for machine translation: Papers from the 1993 AAAI Spring Symposium, March 23-25, Stanford, California. Menlo Park, Ca.: AAAI Press. find
- Alshawi, H. (1992), ed. The Core Language Engine. Cambridge, Mass.: MIT Press. find
- AMTA (1994): Technology Partnerships for Crossing the Language Barrier. Proceedings of the First Conference of the Association for Machine Translation in the Americas, 5-8 October 1994, Columbia, Maryland. find
- Aumaitre, J. M. et al. (1992): TRADEX, un système de traduction de telex. Meta 37 (4), 624-634. find
- Bouillon, P. & Boesefeldt, K. (1992): Problèmes de traduction automatique dans le sous-langage des bulletins d'avalanches. Meta 37(4), 635-646. find
- Boitet, C.(1990): Towards personal MT: general design, dialogue structure, potential role of speech. In: Coling 90 (3), 30-35. find
- Bourbeau, L. et al. (1990): Bilingual generation of weather forecasts in an operations environment. In: Coling 90 (1), 90-92. find
- Brown, P. et al. (1990): A statistical approach to language translation. Computational Linguistics 16, 79-85. find
- Brown, P.F. et al. (1992): Analysis, statistical transfer, and synthesis in machine translation. In: TMI-92, 63-100. find
- Carbonell, J. et al. (1992): JTEC Panel report on machine translation in Japan. Baltimore, MD: Japanese Technology Evaluation Centre. find
- Coling 88: Coling Budapest: Proceedings of the 12th International Conference on Computational Linguistics. Budapest: John von Neumann Society for Computing Sciences, 1988. find
- Coling 90: Coling-90: papers presented to the 13th International Conference on Computational Linguistics...1990 Aug, Helsinki, Finland. Helsinki: Yliopistopaino, 1990. 3 vols. find
- Coling 92: COLING-92: proceedings of the fiftieth International Conference on Computational Linguistics, 23-28/8/1992, Nantes. Grenoble: GETA, 1992. 4 vols. find
- DiMarco, C. (1994): Stylistic choice in machine translation. In: AMTA (1994), 32-39. find
- Dorr, B.J. (1993): Machine translation: a view from the lexicon. Cambridge, Mass.: MIT Press, 1993. find
- Dorr, B.J. & Voss, C.R. (1993): Constraints on the space of MT divergences. In: AAAI (1993), 43-53. 4 - 14 Machine Translation Ten Years On EDR (1990): Proceedings of International Workshop on Electronic Dictionaries, November 8-9, 1990, Oiso, Kanagawa, Japan. Tokyo: EDR. find
- Farwell, D. & Wilks, Y. (1991): ULTRA: a multilingual machine translator. In: MT Summit 3, 19-24. find
- Frederking, R. et al. (1994): Integrating translations from multiple sources within the Pangloss Mark III machine translation system. In: AMTA (1994), 73-80. find
- Funaki, S. (1993): Multi-lingual Machine Translation (MMT) project. In: MT Summit 4, 73-78. find
- Furuse, O. & Iida, H. (1992): Transfer-driven machine translation. In: International Workshop on Fundamental Research for the Future Generation of Natural Language Processing, 30-31 July 1992, Centre for Computational Linguistics, UMIST; 95-111. find
- Gates, D.M. & Shell, P. (1993): Rule-based acquisition and maintenance of lexical and semantic knowledge. In: Sixth Conference of the European Chapter of the Association for Computational Linguistics, 21-23 April 1993, Utrecht University; 149-157. find
- Goodman, K. & Nirenburg, S. (1991), eds.: The KBMT project: a case study in knowledge-based machine translation. San Mateo, Ca.: Morgan Kaufmann. find
- Hansen, V. (1994): PaTrans - a MT system: development and implementation of and experiences from a MT-system. In: AMTA (1994), 114-121. find
- Hutchins, J. (1986): Machine translation: past, present, future. Chichester: Ellis Horwood. find
- Hutchins, J. (1988): Recent developments in machine translation. In: Maxwell, D. et al. (eds.) New directions in machine translation. Conference proceedings, Budapest 18-19 August 1988. Dordrecht: Foris; 7-63. find
- Hutchins, J. (1993): Latest developments in machine translation technology. In: MT Summit 4, 11-34. find
- Iordanskaja, L. et al. (1992): Generation of extended bilingual statistical reports. In: Coling 92 (3), 1019-1023. find
- Isabelle, P. et al. (1988): CRITTER: a translation system for agricultural market reports. In: Coling 88, 261-266. find
- Ishikawa, M. & Sugimura, R. (1992): Natural language analysis using a network model: modification deciding network. In: TMI-92, 55-66. find
- Jacqmin, L.(1992). La traduction automatique au service de l'utilisateur monolingue. Meta 37(4), 610-623. find
- Jensen, K. et al. (1993): Natural language processing: the PLNLP approach. Boston: Kluwer. find
- Kay, M. (1984): Functional unification grammar: a formalism for machine translation. In: Coling 84: 10th International Conference on Computational Linguistics, 2-6 July 1984, Stanford University, California; 75-78. find
- McCord, M. C. (1989): Design of LMT: a Prolog-based machine translation system. Computational Linguistics 15, 33-52. 4 - 15 Machine Translation Ten Years On McLean, I.J. (1992): Example-based machine translation using connectionist matching. In: TMI-92, 35-43. find our card
- Mitkov, R. (1991): Generating public weather reports. In: Proceedings International Conference on Current Issues in Computational Linguistics, 12-14 June 1991, Universiti Sains Malaysia, Penang, Malaysia; 372-378. find
- Mitkov, R. (1992): Discourse-based approach in machine translation. In: International Symposium on Natural Language Understanding and AI, July 12-15, 1992, Kyushu Institute of Technology, Iizuku, Fukuoka; 225-230. find
- Miura, M. et al. (1992): Learning mechanism in machine translation system "PIVOT". In: Coling 92, (2), 693-699. find
- Morimoto, T. & Kurematsu, A. (1993): Automatic speech translation at ATR. In: MT Summit 4, 83-96 find
- MT Summit 3: MT Summit III, July 1-4 1991, Washington D.C., USA. find
- MT Summit 4: MT Summit IV: International Co-operation for Global Communication, July 20-22, 1993, Kobe, Japan. find
- Nagao, M. (1984): A framework of a mechanical translation between Japanese and English by analogy principle. In: Elithorn, A. & Banerji,R.(eds.) Artificial and human intelligence. Amsterdam: North-Holland; 173-180. find
- Neff, M.S. & McCord, M.C. (1990): Acquiring lexical data from machine-readable dictionary resources for machine translation. In: TMI-90, 85-90. find
- Nirenburg, S. et al. (1992): Machine translation: a knowledge-based approach. San Mateo, Ca.: Morgan Kaufmann. find
- Nishida, F. & Takematsu, S. (1990): Automated procedures for the improvement of a machine translation system by feedback from postediting. Machine Translation 5(3), 223-246. find
- Okumura, A. et al. (1992): A pattern-learning based hybrid model for the syntactic analysis of structural relationships among Japanese clauses. In: TMI-92, 45-54. find
- Rimon, M. et al.(1991): Advances in machine translation research in IBM. In: MT Summit 3, 11-18. find
- Sadler, V. (1989): Working with analogical semantics: disambiguation techniques in DLT. Dordrecht: Foris. find
- Sharp, R. & Streiter, O. (1992): Simplifying the complexity of machine translation. Meta 37 (4), 681-692. find
- Somers, H.L. (1992): Interactive multilingual text generation for a monolingual user. In: TMI-92, 151-161. find
- Takebayashi, Y. (1993): EDR electronic dictionary. In: MT Summit 4, 117-126 find
- Takeda, K. et al. (1992): Shalt2 - a symmetric machine translation system with conceptual transfer. In: Coling 92 (3), 1034-1038. 4 - 16 Machine Translation Ten Years On TMI-90: Third International Conference on Theoretical and Methodological Issues in Machine Translation of Natural Languages, 11-13 June 1990, Austin, TX, USA. find
- TMI-92: Fourth International Conference on Theoretical and Methodological Issues in Machine Translation of Natural Languages. Empricist vs Rational Methods in MT, June 25-27, 1992, Montreal, Canada. find
- TMI-93: Fifth International Conference on Theoretical and Methodological Issues in Machine Translation of Natural Languages. MT in the Next Generation, July 14-16, 1993, Kyoto, Japan. find
- Van der Steen, G. & Dijenborgh, B.J.(1992): On-line correction and translation of industrial texts. In: Translating and the computer 14 (London: ASLIB), 135-164. find
- Vasconcellos, M. (1993): The present state of machine translation usage technology, or: How do I use thee? Let me count the ways! In: MT Summit 4, 35-46. Also in: MT News International 6 (September 1993), 12-17. find
- Wahlster, W. (1993): Verbmobil: translation of face-to-face dialogs. In: MT Summit 4, 127-135 find
- Whitelock, P. (1992): Shake-and-bake translation. In: Coling 92 (2), 784-791. find our card
- Woszczyna, M. et al. (1993): Recent advances in JANUS: a speech translation system. In: TMI-93, 195-200. find
- Zaki Abu Bakar, A. et al.(1991): Malay official letters translation. In: Proceedings International Conference on Current Issues in Computational Linguistics, 12-14 June 1991, Universiti Sains Malaysia, Penang, Malaysia; 413-420. 4 - 17 Machine Translation Ten Years On find
This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/hutchins-1994-research-methods-system-designs-machine/