“A Neural Attention Model for Abstractive Sentence Summarization”, Alexander M. Rush, Sumit Chopra, Jason Weston2015-09-02 (CNN, MLP NN)⁠:

Summarization based on text extraction is inherently limited, but generation-style abstractive methods have proven challenging to build.

In this work, we propose a fully data-driven approach to abstractive sentence summarization. Our method uses a local attention-based model that generates each word of the summary conditioned on the input sentence.

While the model is structurally simple [cf. Bengio et al 2003], it can easily be trained end-to-end and scales to a large amount of training data.

The model shows performance gains on the DUC-2004 shared task compared with several strong baselines.