|
The Tenjinno Machine Translation Competition | |||||||||||
|
Task descriptionTenjinno is a competition that combines grammatical inference with machine translation. The task is to create a machine translation system based on the data only. This data is generated automatically. How this is done (any why) can be found on the background page. Essentially, we have generated an automatic machine translation system that is based on finite state transduces and syntax directed translation schemas. Based on these formal models, we have generated sentences in the source language and corresponding translations in the target language. The test data, that can be used to train you system, contain both the source and target sentences, but the test data only contains the source sentences. The task is to generate the corresponding target sentences. These can be submitted on the oracle page. The oracle will test the translated sentences for correctness and will tell you if the result is good enough. To win this competition you need to do the following:
|
|||||||||||
| For any comments or questions about these pages please contact the Tenjinno organisers. |
|
Brad Starkie Menno van Zaanen Dominique Estival |