Comparing Computational Architectures for Automated Journalism
This addresses the problem of automated journalism for Brazilian Portuguese content creators, though it appears incremental as it compares existing architectures rather than proposing new ones.
This study compared template-based/pipeline-based architectures against neural end-to-end approaches for generating Brazilian Portuguese texts from structured data, finding that explicit intermediate steps produced better texts with less data hallucination and better generalization to unseen inputs.
The majority of NLG systems have been designed following either a template-based or a pipeline-based architecture. Recent neural models for data-to-text generation have been proposed with an end-to-end deep learning flavor, which handles non-linguistic input in natural language without explicit intermediary representations. This study compares the most often employed methods for generating Brazilian Portuguese texts from structured data. Results suggest that explicit intermediate steps in the generation process produce better texts than the ones generated by neural end-to-end architectures, avoiding data hallucination while better generalizing to unseen inputs. Code and corpus are publicly available.