Countering the Effects of Lead Bias in News Summarization via Multi-Stage Training and Auxiliary Losses
arXiv:1909.04028
Abstract
Sentence position is a strong feature for news summarization, since the lead often (but not always) summarizes the key points of the article. In this paper, we show that recent neural systems excessively exploit this trend, which although powerful for many inputs, is also detrimental when summarizing documents where important content should be extracted from later parts of the article. We propose two techniques to make systems sensitive to the importance of content in different parts of the article. The first technique employs 'unbiased' data; i.e., randomly shuffled sentences of the source document, to pretrain the model. The second technique uses an auxiliary ROUGE-based loss that encourages the model to distribute importance scores throughout a document by mimicking sentence-level ROUGE scores on the training data. We show that these techniques significantly improve the performance of a competitive reinforcement learning based extractive system, with the auxiliary loss being more powerful than pretraining.
arXiv comment5 pages, accepted at EMNLP 2019
References in corpus (7)
- Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
- Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
- Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor
- Teaching Machines to Read and Comprehend
- A Deep Reinforced Model for Abstractive Summarization
- SummaRuNNer: A Recurrent Neural Network based Sequence Model for Extractive Summarization of Documents
- Tune: A Research Platform for Distributed Model Selection and Training