Data-to-text Generation with Variational Sequential Planning
This addresses the problem of producing coherent multi-paragraph documents from structured data for applications like automated reporting, though it appears incremental as it builds on existing planning methods.
The paper tackled generating long-form text from non-linguistic data by proposing a neural model with a variational sequential planning component, resulting in outperforming strong baselines on benchmarks like RotoWire and MLB and showing sample efficiency with limited training data, such as a few hundred instances.
We consider the task of data-to-text generation, which aims to create textual output from non-linguistic input. We focus on generating long-form text, i.e., documents with multiple paragraphs, and propose a neural model enhanced with a planning component responsible for organizing high-level information in a coherent and meaningful way. We infer latent plans sequentially with a structured variational model, while interleaving the steps of planning and generation. Text is generated by conditioning on previous variational decisions and previously generated text. Experiments on two data-to-text benchmarks (RotoWire and MLB) show that our model outperforms strong baselines and is sample efficient in the face of limited training data (e.g., a few hundred instances).