Framework for Abstractive Summarization using Text-to-Text Generation

摘要: We propose a new, ambitious framework for abstractive summarization, which aims at selecting the content of summary not from sentences, but an abstract representation source documents. This relies on concept Information Items (InIt), we define as smallest element coherent information in text or sentence. Our differs previous summarization models requiring semantic analysis text. present first attempt made developing system this framework, along with evaluation results it TAC 2010. also related work, both within and outside automatic domain.