Global-aware Beam Search for Neural Abstractive Summarization

This study develops a calibrated beam-based algorithm with global awareness for neural abstractive summarization, aiming to improve the local optimality problem of the original beam search in a rigorous way. Specifically, a novel global protocol is proposed based on the attention distribution to stipulate how a global optimal hypothesis should attend to the source. A global scoring function is then developed to regulate beam search to generate summaries in a more near-global optimal fashion. This novel design enjoys a distinctive property, i.e. the global attention distribution could be predicted before inference, enabling stepwise improvements on the beam search through the global scoring function. Extensive experiments on $9$ datasets show that the global-aware inference significantly improves state-of-the-art summarization models even using empirical hyper-parameters. The algorithm is also proven robust as it remains to generate meaningful texts with corrupted attention distributions. The codes and a comprehensive set of examples are available.