Machine Translation Using Isomorphic UCGs

What it is about

The paper applies Unification Categorial Grammar (UCG), a lexicalist formalism in which combinatory information sits in lexical items, to the isomorphic-grammar method of Landsbergen's Rosetta project. Source and target grammars are written in parallel so that translation equivalents have isomorphic derivations. Semantic information enters as constraints on the translation relation, not as a level of text representation. The authors show an implemented bidirectional English to Spanish fragment, and they list open problems: the freer word order of Spanish, the characterisation of the basic bilingual signs (the semantics must unify, but that is not sufficient), and efficiency where bilingual signs need discontinuous or null constituents. The paper gives no figures.

Ideas

• I1. Word order as a counted rule: for a language with freer order, learn the order of the meta tokens from the aligned verses, with the classes of the Markov chains in en/src/lin.rs, instead of writing the order by hand. The check is the exact-match rate on verses with a subject-object swap, twice for determinism.
• I2. Empty chunks: allow a meta token with no surface word, such as a pronoun that a Hebrew verb form carries. Count, from the Hebrew to English pairs, how many meta tokens need an empty chunk and where they fall.
• I3. A search budget: cap the number of candidate chunks per meta token in the encoder, and measure the time against the exact-match rate. The paper's efficiency problem becomes a curve, not an argument.

For Mova

The code is en/src/lin.rs (word order from the dependency tree) and tools/translate/src/profile.rs (chunk choice). The word-order problem is open problem 7 for the Hebrew reader, and I1 is a concrete way to count it.

Topics: Machine translation

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.3115/991635.991642

@inproceedings{beaven1988machine,
  title = {{Machine Translation Using Isomorphic UCGs}},
  author = {Beaven, John L. and Whitelock, Pete},
  booktitle = {Coling Budapest 1988 Volume 1: International Conference on Computational Linguistics},
  year = {1988},
  doi = {10.3115/991635.991642},
  url = {https://doi.org/10.3115/991635.991642}
}

References

11 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Calder, J., Moens, M. and Zeevat, H. (1986) A UCG interpreter. ACORD Deliverable T2.6. Centre for Cognitive Science, Edinburgh University. find
2. Johnson, M. and Klein, E. (1986) Discourse, ~naphora and parsing. In Proceedings of the 11th International Conference on Computational Linguistics and the 24th Annual Meeting of the Association for Computational Liv~,guistics, Institut lure Kommunikationsforschung und Phonetik~ Bonn University, Boon, August 1986. find
3. I/amp, It. (1981) A theory of truth and seroantic representation. In Groenendijk, J.A.G., Janssen, T.M.V. and Stokhof, M.B.J. (Eds) Formal Methods in the Study of Language, Vo1136, pp 227-322. Amsterdam: Mathematical Centre 'lh'acts. find
4. Karttunen, L. (1986) Radical Lexicalism. CSLI-86-68, Centre for the Study of Language and Information, Stanford University, California. find
5. Landsbergen, J. (1987). Isomorphic Grammars and their Use in the ROSETTA Translation System. In King, M. (Ed) Machine Translation Today: the State of the Art. Pro-ceedi~gs of the Third Lugano Tutorial, Lugano, Switzerland, 2--7 April 1984. Edinburgh University Press. find
6. Landsbergen, J. (1987) Montague Grammar and Machine 'l.k'anslatioa. In Whitelock et aL (Eds). Liguistic Theorz and Computer Applications. Academic Press. find
7. Pollard, C. (1985). Lectures on tlPSG. Unpublished lecture l~otes, GSLI, Stanford University. find
8. Shieber, S. (1986) An Introduction to Unification-based Ap~,roaehes to Grammar. Lecture Notes Number 4. Center for tim Study of Language and Information, Stanford University. find
9. Whitelock, P. (1988) The Organisation of a Bilingual Lexi,~on. DAI Working Paper, Dept. of Artificial Intelligence, "(hfiq. of Edinburgh. find
10. 7.eevat, H. (1986). A specification oflnL. Unpublished Internal ACORD Report. Centre for Cognitive Science, University of Edinburgh. find
11. Zeevat, H., Klein, E., and Calder, J. (1987). Unification Categorial Grammar. In Haddock, N., Klein, E. and Moro rill, G. (Eds) (1987). Working Papers in Cognitive Science, Vo|. 1: Categodal Grammar, Unification Grammar and Parsing. Centre for Cognitive Science, University of Edinburgh. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/beaven-1988-machine-translation-isomorphic-ucgs/