A Deep Reinforced Model for Abstractive Summarization
Romain Paulus, Caiming Xiong、Richard Socher Salesforce Research 172 University Avenue Palo Alto, CA 94301, USA
Abstract
Attentional, RNN-based encoder-decoder models for abstractive summarization have achieved good performance on short input and output sequences. For longer documents and summaries however these models often include repetitive and incoherent phrases. We introduce a neural network model with a novel intra-attention that attends over the input and continuously generated output separately, and a new training method that combines standard supervised word prediction and reinforcement learning (RL). Models trained only with supervised learning often exhibit “exposure bias” – they assume ground truth is provided at each step during training. However, when standard word prediction is combined with the global sequence prediction training of RL the resulting summaries become more readable. We evaluate this model on the CNN/Daily Mail and New York Times datasets. Our model obtains a 41.16 ROUGE-1 score on the CNN/Daily Mail dataset, an improvement over previous state-of-the-art models. Human evaluation also shows that our model produces higher quality summaries.
中文速览
针对长文档的抽象式摘要生成（abstractive summarization）任务，现有基于注意力机制的编码器-解码器模型容易产生重复短语和语义不连贯的问题，本文提出了一种结合"内部注意力"（intra-attention）与混合训练目标的神经网络模型来解决这一难题。具体做法是：在编码端引入时序内部注意力，记录历史注意力权重以避免重复关注相同输入片段；在解码端引入解码器内部注意力，让模型在生成每个词时能回顾已生成的内容；同时将传统的有监督最大似然训练与强化学习（reinforcement learning）中的自我批评策略梯度算法相结合，缓解训练与测试阶段的"暴露偏差"（exposure bias）问题。实验结果表明，该模型在CNN/Daily Mail数据集上取得了41.16的ROUGE-1分数，超越此前最优方法，并在人工评估中也获得了更高的可读性评分。这项工作不仅推动了长文档摘要生成的技术上限，也为如何将强化学习与有监督学习相融合以提升生成质量提供了有价值的示范。
原文 arXiv:1705.04304;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1705.04304v3