Generating Wikipedia by Summarizing Long Sequences
This addresses the problem of scalable multi-document summarization for creating encyclopedic content, representing a novel method for a known bottleneck.
The paper tackled generating Wikipedia articles by summarizing long source documents, using extractive summarization to identify key information and a novel decoder-only neural model for abstractive generation, resulting in fluent articles with improved perplexity, ROUGE scores, and human evaluations.
We show that generating English Wikipedia articles can be approached as a multi- document summarization of source documents. We use extractive summarization to coarsely identify salient information and a neural abstractive model to generate the article. For the abstractive model, we introduce a decoder-only architecture that can scalably attend to very long sequences, much longer than typical encoder- decoder architectures used in sequence transduction. We show that this model can generate fluent, coherent multi-sentence paragraphs and even whole Wikipedia articles. When given reference documents, we show it can extract relevant factual information as reflected in perplexity, ROUGE scores and human evaluations.