NLG System Research Articles

Decoding Methods in Neural Language Generation: A Survey

Neural encoder-decoder models for language generation can be trained to predict words directly from linguistic or non-linguistic inputs. When generating with these so-called end-to-end models, however, the NLG system needs an additional decoding procedure that determines the output sequence, given the infinite search space over potential sequences that could be generated with the given vocabulary. This survey paper provides an overview of the different ways of implementing decoding on top of neural network-based generation models. Research into decoding has become a real trend in the area of neural language generation, and numerous recent papers have shown that the choice of decoding method has a considerable impact on the quality and various linguistic properties of the generation output of a neural NLG system. This survey aims to contribute to a more systematic understanding of decoding methods across different areas of neural NLG. We group the reviewed methods with respect to the broad type of objective that they optimize in the generation of the sequence—likelihood, diversity, and task-specific linguistic constraints or goals—and discuss their respective strengths and weaknesses.

Automatic generation of weather forecast texts using comprehensive probabilistic generation-space models

AbstractTwo important recent trends in natural language generation are (i) probabilistic techniques and (ii) comprehensive approaches that move away from traditional strictly modular and sequential models. This paper reports experiments in whichpcru– a generation framework that combines probabilistic generation methodology with a comprehensive model of the generation space – was used to semi-automatically create five different versions of a weather forecast generator. The generators were evaluated in terms of output quality, development time and computational efficiency against (i) human forecasters, (ii) a traditional handcrafted pipelinednlgsystem and (iii) ahalogen-style statistical generator. The most striking result is that despite acquiring all decision-making abilities automatically, the bestpcrugenerators produce outputs of high enough quality to be scored more highly by human judges than forecasts written by experts.

Lessons from a failure: Generating tailored smoking cessation letters

stop is a Natural Language Generation ( nlg) system that generates short tailored smoking cessation letters, based on responses to a four-page smoking questionnaire. A clinical trial with 2553 smokers showed that stop was not effective; that is, recipients of a non-tailored letter were as likely to stop smoking as recipients of a tailored letter. In this paper we describe the stop system and clinical trial. Although it is rare for ai papers to present negative results, we believe that useful lessons can be learned from stop. We also believe that the ai community as a whole could benefit from considering the issue of how, when, and why negative results should be reported; certainly a major difference between ai and more established fields such as medicine is that very few ai papers report negative results.