Robust Neural Abstractive Summarization Systems and Evaluation against Adversarial Information
arXiv:1810.06065
Abstract
Sequence-to-sequence (seq2seq) neural models have been actively investigated for abstractive summarization. Nevertheless, existing neural abstractive systems frequently generate factually incorrect summaries and are vulnerable to adversarial information, suggesting a crucial lack of semantic understanding. In this paper, we propose a novel semantic-aware neural abstractive summarization model that learns to generate high quality summaries through semantic interpretation over salient content. A novel evaluation scheme with adversarial samples is introduced to measure how well a model identifies off-topic information, where our model yields significantly better performance than the popular pointer-generator summarizer. Human evaluation also confirms that our system summaries are uniformly more informative and faithful as well as less redundant than the seq2seq model.
References in corpus (12)
- Neural Machine Translation by Jointly Learning to Align and Translate
- Sequence to Sequence Learning with Neural Networks
- Teaching Machines to Read and Comprehend
- A Deep Reinforced Model for Abstractive Summarization
- Incorporating Copying Mechanism in Sequence-to-Sequence Learning
- Faithful to the Original: Fact Aware Neural Abstractive Summarization
- Fast Abstractive Summarization with Reinforce-Selected Sentence Rewriting
- Stochastic Language Generation in Dialogue using Recurrent Neural Networks with Convolutional Sentence Reranking
- Multi-task Sequence to Sequence Learning
- Temporal Attention Model for Neural Machine Translation
- Deep Multitask Learning for Semantic Dependency Parsing
- Neural Network-Based Abstract Generation for Opinions and Arguments
Cited by in corpus (7)
- An Empirical Survey of Data Augmentation for Limited Data Learning in NLP
- BIGPATENT: A Large-Scale Dataset for Abstractive and Coherent Summarization
- An Entity-Driven Framework for Abstractive Summarization
- Point-less: More Abstractive Summarization with Pointer-Generator Networks
- Sensitivity of BLANC to human-scored qualities of text summaries
- Augmented Abstractive Summarization With Document-LevelSemantic Graph
- Estimation of Summary-to-Text Inconsistency by Mismatched Embeddings