Copy mechanism and tailored training for character-based data-to-text generation

26 Apr 2019  ·  Marco Roberti, Giovanni Bonetta, Rossella Cancelliere, Patrick Gallinari ·

In the last few years, many different methods have been focusing on using deep recurrent neural networks for natural language generation. The most widely used sequence-to-sequence neural methods are word-based: as such, they need a pre-processing step called delexicalization (conversely, relexicalization) to deal with uncommon or unknown words. These forms of processing, however, give rise to models that depend on the vocabulary used and are not completely neural. In this work, we present an end-to-end sequence-to-sequence model with attention mechanism which reads and generates at a character level, no longer requiring delexicalization, tokenization, nor even lowercasing. Moreover, since characters constitute the common "building blocks" of every text, it also allows a more general approach to text generation, enabling the possibility to exploit transfer learning for training. These skills are obtained thanks to two major features: (i) the possibility to alternate between the standard generation mechanism and a copy one, which allows to directly copy input facts to produce outputs, and (ii) the use of an original training pipeline that further improves the quality of the generated texts. We also introduce a new dataset called E2E+, designed to highlight the copying capabilities of character-based models, that is a modified version of the well-known E2E dataset used in the E2E Challenge. We tested our model according to five broadly accepted metrics (including the widely used BLEU), showing that it yields competitive performance with respect to both character-based and word-based approaches.

PDF Abstract

Datasets


Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Benchmark
Data-to-Text Generation E2E NLG Challenge EDA_CS (TL) BLEU 65.80 # 6
NIST 8.5615 # 4
METEOR 45.16 # 5
ROUGE-L 67.40 # 9
CIDEr 2.1803 # 6
Data-to-Text Generation E2E NLG Challenge EDA_CS BLEU 67.05 # 2
NIST 8.5150 # 5
METEOR 44.49 # 9
ROUGE-L 68.94 # 4
CIDEr 2.2355 # 3

Methods


No methods listed for this paper. Add relevant methods here