Attentive Natural Language Generation from Abstract Meaning Representation

Radha Senthilkumar and S. Afrish Khan
Additional contact information
Radha Senthilkumar: MIT, Anna University, Information Technology
S. Afrish Khan: MIT, Anna University, Information Technology

A chapter in New Trends in Computational Vision and Bio-inspired Computing, 2020, pp 1649-1657 from Springer

Abstract: Abstract Natural Language Generation takes a key role in presenting data as text or speech. The translation into a natural language from semantic representation is similar to Neural Machine Translation. We use a similar methodology known as Seq2Seq modelling for generating natural language. The usage of common semantic representation such as Abstract Meaning Representation allows adding naturalness to the generated sentences while being domain neutral. Recurrent Neural Network based autoencoder learns a hidden representation from semantic input which is then used to generate natural language. Long-Short Term Memory while in theory being capable of learning long-term dependencies fails to capture the correct information required for generation. We introduce attention mechanism as a resolution to improve capturing contextually important information. The resulting model has a significantly improved accuracy.

Keywords: Natural languages; Computational linguistics; Machine learning (search for similar items in EconPapers)
Date: 2020
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:sprchp:978-3-030-41862-5_169

Ordering information: This item can be ordered from
http://www.springer.com/9783030418625

DOI: 10.1007/978-3-030-41862-5_169

Access Statistics for this chapter

More chapters in Springer Books from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().