Data-Driven News Generation for Automated Journalism

Abstract

Despite increasing amounts of data and ever improving natural language generation techniques, work on automated journalism is still relatively scarce. In this paper, we explore the field and challenges associated with building a journalistic natural language generation system. We present a set of requirements that should guide system design, including transparency, accuracy, modifiability and transferability. Guided by the requirements, we present a data-driven architecture for automated journalism that is largely domain and language independent. We illustrate its practical application in the production of news articles upon a user request about the 2017 Finnish municipal elections in three languages, demonstrating the successfulness of the data-driven, modular approach of the design. We then draw some lessons for future automated journalism.

Personal Research Assistant for Online Exploration of Historical News

“…The PRA produces a set of natural language reports detailing its findings. These are produced by an automatic natural language generation system [1] and can be generated in English, French, Finnish or German.…”

Towards Data-Driven Generation of Visualizations for Automatically Generated News Articles

“…The visualization generator is developed as a complementary component to an NLG system that automatically generates news articles from structured data based on a user query. This system's architecture is based on the architecture presented by Leppänen et al [13], with added visualization components. Changes were made to the natural language generation components as well, but these are not relevant in terms of the contribution of the present paper.…”

“…Note that while the caption itself is obviously domain-specific, the generator for captions is not: all the domain-specific information is given as parameters in the form of a domain-specific lexicon. Adapted from [13].…”

“…In this work, we investigate the challenge of automatic generation of visuals to accompany textual news articles. These news articles are themselves produced automatically using a natural language generation (NLG) system [13] that takes as its input structured, statistical data and outputs textual stories that describe the most newsworthy aspects of the dataset.…”

Neural data-to-text generation: A comparison between pipeline and end-to-end architectures

“…Data-to-text Natural Language Generation (NLG) is the computational process of generating meaningful and coherent natural language text to describe non-linguistic input data (Gatt and Krahmer, 2018). Practical applications can be found in domains such as weather forecasts (Mei et al, 2016), health care (Portet et al, 2009), feedback for car drivers (Braun et al, 2018), diet management (Anselma and Mazzei, 2018), election results (Leppänen et al, 2017) and sportscasting news (van der Lee et al, 2017).…”