GGP: A Graph-based Grouping Planner for Explicit Control of Long Text Generation
This addresses the challenge of controllable long-text generation for applications such as commercial advertising, though it appears incremental as it builds on the first-plan-then-generate idea.
The paper tackles the problem of generating illogical and uncontrollable long texts in scenarios like story or advertising generation by proposing a graph-based grouping planner (GGP) that first plans then generates, resulting in significant performance improvements over baselines on three datasets.
Existing data-driven methods can well handle short text generation. However, when applied to the long-text generation scenarios such as story generation or advertising text generation in the commercial scenario, these methods may generate illogical and uncontrollable texts. To address these aforementioned issues, we propose a graph-based grouping planner(GGP) following the idea of first-plan-then-generate. Specifically, given a collection of key phrases, GGP firstly encodes these phrases into an instance-level sequential representation and a corpus-level graph-based representation separately. With these two synergic representations, we then regroup these phrases into a fine-grained plan, based on which we generate the final long text. We conduct our experiments on three long text generation datasets and the experimental results reveal that GGP significantly outperforms baselines, which proves that GGP can control the long text generation by knowing how to say and in what order.