Automatic Document Sketching: Generating Drafts from Analogous Texts
This addresses the challenge for writers needing efficient document drafting, but it is incremental as it builds on existing language models and datasets.
The paper tackles the problem of generating entire draft documents from sets of analogous texts, introducing a new task called document sketching to overcome limitations of language models in global editing. They report experiments using a Wikipedia-based dataset and methods like transformer-based mixture of experts with reinforcement learning, evaluated through automated and human assessments.
The advent of large pre-trained language models has made it possible to make high-quality predictions on how to add or change a sentence in a document. However, the high branching factor inherent to text generation impedes the ability of even the strongest language models to offer useful editing suggestions at a more global or document level. We introduce a new task, document sketching, which involves generating entire draft documents for the writer to review and revise. These drafts are built from sets of documents that overlap in form - sharing large segments of potentially reusable text - while diverging in content. To support this task, we introduce a Wikipedia-based dataset of analogous documents and investigate the application of weakly supervised methods, including use of a transformer-based mixture of experts, together with reinforcement learning. We report experiments using automated and human evaluation methods and discuss relative merits of these models.