Training a Natural Language Generator From Unaligned Data

By Ondrej Dusek, Filip Jurcícek. 2015

46 citations recorded by the source corpus.

We present a novel syntax-based natural language generation system that is trainable from unaligned pairs of input meaning representations and output sentences. It is divided into sentence planning, which incrementally builds deep-syntactic dependency trees, and surface realization. Sentence planner is based on A* search with a perceptron ranker that uses novel differing subtree updates and a simple future promise estimation; surface realization uses a rule-based pipeline from the Treex NLP toolkit. Our first results show that training from unaligned data is feasible, the outputs of our generator are mostly fluent and relevant.

Subfields: Artificial Intelligence, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

Citations in the corpus

The E2E Dataset: New Challenges For End-to-End Generation (2017, background): “…promise rapid development of NLG components in new domains: They jointly learn sentence planning and surface realisation from non-aligned data (Duˇsek and Jurˇc´ıˇcek, 2015; Wen et al., 2015; Mei et al., 2016; Wen et al., 2016; Sharma et al., 2016; Duˇsek and Jurˇc´ıˇcek, 2016a; Lampouras and…”
Sequence-to-Sequence Generation for Spoken Dialogue via Deep Syntax Trees and Strings (2016, methodology): “ited domain (Dušek and Jurčı́ček, 2015), leaving the seq2seq generator as the major error source.”
Survey of the State of the Art in Natural Language Generation: Core tasks, applications and evaluation (2017, background): “For example, Dušek and Jurč́ıček (2015) use the dialogue acts from the bagel dataset (Mairesse et al.”
Findings of the E2E NLG Challenge (2018, background): “(Dušek and Jurčı́ček, 2015; Wen et al., 2015b; Mei et al., 2016; Wen et al., 2016; Sharma et al., 2016; Dušek and Jurčı́ček, 2016a; Lampouras and Vlachos, 2016).”
Neural Belief Tracker: Data-Driven Dialogue State Tracking (2016, background): “, 2016a,b); the system action is then verbalised by the natural language generator (Wen et al., 2015a,b; Dušek and Jurčı́ček, 2015).”
New Language Pairs in TectoMT (2015, methodology): “Besides the Czech-to-English translation, it is used in other TectoMT systems translating into English within the QTLeap project and in the TGen natural language generator (Dušek and Jurˇcíˇcek, 2015).”
Imitation learning for language generation from unaligned data (2016, background): “(2015) and Dušek and Jurcı́cek (2015) generated in their work using their best configurations.”
Handling Rare Items in Data-to-Text Generation (2018, background): “To address these issues, neural approaches to data-to-text generation typically resort either to delexi-calisation (Wen et al., 2015; Duˇsek and Jurcicek, 2015; Trisedya et al., 2018; Chen et al., 2018) or to a copy mechanism (Chen, 2018; Elder et al., 2018; Gehrmann et al., 2018).”
The aNALoGuE Challenge: Non Aligned Language GEneration (2016, methodology): “This shared task aims to strengthen recent research on corpus-based NLG from unaligned data, e.g. (Duˇsek and Jurcicek, 2015; Wen et al., 2015; Mei et al., 2015; Sharma et al., 2016).”