Prenominal Modifier Ordering via Multiple Sequence Alignment

Producing a fluent ordering for a set of prenominal modifiers in a noun phrase (NP) is a problematic task for natural language generation and machine translation systems. We present a novel approach to this issue, adapting multiple sequence alignment techniques used in computational biology to the alignment of modifiers. We describe two training techniques to create such alignments based on raw text, and demonstrate ordering accuracies superior to earlier reported approaches.

• Type: Journal
• Year: 2010
• Where: NAACL
• Authors: Aaron Dunlop, Margaret Mitchell, Brian Roark
• Keywords: Computational Linguistics, Machine Translation Systems, NAACL 2010, Natural Language Generation, Sequence Alignment Techniques