We describe the CMU systems submitted to the 2013 WMT shared task in machine translation. We participated in three language pairs, French–English, Russian– English, and English–Russian. Our particular innovations include: a labelcoarsening scheme for syntactic tree-totree translation and the use of specialized modules to create “synthetic translation options” that can both generalize beyond wha...

Chinese and Spanish are the most spoken languages in the world. However, there is not much research done in machine translation for this language pair. We experiment with the parallel Chinese-Spanish corpus (United Nations) to explore alternatives of SMT strategies which consist on using a pivot language. Particularly, two well-known alternatives are shown for pivoting: the cascade system and t...

Connectionist Models can be considered as an encouraging approach to Example-Based Machine Translation. However, the neural translators developed in the literature are quite complex and require great human effort to classify and prepare training data. This paper presents an effective and more simple text-to-text connectionist translator with which translations from the source to the target lang...

The University of Exeter participated in the CLEF 2001 bilingual task. The main objectives of our experiments were to compare retrieval performance for different topic languages with similar easily available machine translation resources and to explore the application of new pseudo relevance feedback techniques recently developed at Exeter to Cross-Language Information Retrieval (CLIR). We also...

In this paper the UHH submission to the WMT17 Metrics Shared Task is presented, which is based on sequence and tree kernel functions applied to the reference and candidate translations. In addition we also explore the effect of applying the kernel functions on the source sentence and a back-translation of the MT output, but also on the pair composed of the candidate translation and a pseudo-ref...

in this paper, some results of the chebyshev type integral inequality for the pseudo-integral are proven. the obtained results, are related to the measure of a level set of the maximum and the sum of two non-negative integrable functions. finally, we applied our results  to the case of comonotone functions.

Human intelligence exhibits compositional generalization (i.e., the capacity to understand and produce unseen combinations of seen components), but current neural seq2seq models lack such ability. In this paper, we revisit iterative back-translation, a simple yet effective semi-supervised method, investigate whether how it can improve generalization. work: (1) We first empirically show that bac...

This paper reports experimental results of cross-language information retrieval (CLIR) from German to French, in which a hybrid approach of query and document translation was attempted, i.e, combining results of query translation (German to French) and of document translation (French to German). In order to avoid too high complexity of computation for translating a large amount of texts in docu...