Papers by Florie Lambrey
GenDR: A Generic Deep Realizer with Complex Lexicalization (L18-1)
Copied to clipboard
| Challenge: | Generic deep realizers are used for natural language generation, but they are not yet fully dominated by statistical or neuronal methods. |
| Approach: | They propose a generic deep realizer that produces syntactic dependency structures in languages . they use a graph transducer to lexicalize multiword expressions and build on it . |
| Outcome: | The proposed system produces syntactic dependency structures in English, French, Lithuanian and Persian . it is generic in that it is designed to operate across a wide range of languages and applications . |