INSET: Sentence Infilling with INter-SEntential Transformer
This addresses the problem of document auto-completion and meeting note expansion for natural language generation applications, but it is incremental as it builds on existing pre-trained models.
The paper tackles the task of missing sentence generation (sentence infilling) by proposing a framework that decouples understanding, discourse-level planning, and generation, leveraging pre-trained models like BERT and GPT-2. It demonstrates effectiveness in learning sentence representations and generating context-fitting missing sentences.
Missing sentence generation (or sentence infilling) fosters a wide range of applications in natural language generation, such as document auto-completion and meeting note expansion. This task asks the model to generate intermediate missing sentences that can syntactically and semantically bridge the surrounding context. Solving the sentence infilling task requires techniques in natural language processing ranging from understanding to discourse-level planning to generation. In this paper, we propose a framework to decouple the challenge and address these three aspects respectively, leveraging the power of existing large-scale pre-trained models such as BERT and GPT-2. We empirically demonstrate the effectiveness of our model in learning a sentence representation for generation and further generating a missing sentence that fits the context.