SimpleNLG-ZH: a Linguistic Realisation Engine for Mandarin

Abstract

We introduce SimpleNLG-ZH, a realisation engine for Mandarin that follows the software design paradigm of SimpleNLG (Gatt and Reiter, 2009). We explain the core grammar (morphology and syntax) and the lexicon of SimpleNLG-ZH, which is very different from English and other languages for which SimpleNLG engines have been built. The system was evaluated by regenerating expressions from a body of test sentences and a corpus of humanauthored expressions. Human evaluation was conducted to estimate the quality of regenerated sentences.

Arabic NLG Language Functions

“…However, SimpleNLG is still considered the most popular method since it was presented by Gatt and Reiter (2009), It is a rulebased realiser that uses three main steps syntactic, morphological and orthographic realisations to produce a correctly inflected and high-quality output. It was adapted to many languages includ-ing (in chronological order) French (Vaudry andLapalme, 2013), Italian (Mazzei et al, 2016), Spanish (Ramos-Soto et al, 2017), Galician (Cascallar-Fuentes et al, 2018), Dutch (de Jong andTheune, 2018), Mandarin (Chen et al, 2018) and German (Braun et al, 2019).…”

SimpleNLG-TI: Adapting SimpleNLG to Tibetan

“…To realise well-formed text, SimpleNLG-EN adopts rule-based approach based on lexicon and rules with rich linguistic knowledge. The paradigm of Sim-pleNLG has been successfully implemented into many natural language such as Chinese SimpleNLG-ZH (Chen et al, 2018) and German SimpleNLG-DE (Braun et al, 2019), and SimpleNLG-EnFr (Vaudry and Lapalme, 2013).…”

Neural Generation for Czech: Data and Baselines

“…There have been a few approaches using handcrafted grammars (Bateman, 1997;Allman et al, 2012). The procedural Sim-pleNLG realizer (Gatt and Reiter, 2009) has also been ported into multiple languages (Bollmann, 2011;Vaudry and Lapalme, 2013;de Oliveira and Sripada, 2014;Mazzei et al, 2016;Ramos-Soto et al, 2017;Cascallar-Fuentes et al, 2018;Chen et al, 2018;de Jong and Theune, 2018). Further works using multilingual rule-based surface realiza-tion pipelines were developed in the context of machine translation (Aikawa et al, 2001;Žabokrtský et al, 2008;Dušek et al, 2015).…”