Sample Efficient Text Summarization Using a Single Pre-Trained Transformer

Abstract

Language model (LM) pre-training has resulted in impressive performance and sample efficiency on a variety of language understanding tasks. However, it remains unclear how to best use pre-trained LMs for generation tasks such as abstractive summarization, particularly to enhance sample efficiency. In these sequence-to-sequence settings, prior work has experimented with loading pre-trained weights into the encoder and/or decoder networks, but used non-pre-trained encoder-decoder attention weights. We instead use a pre-trained decoder-only network, where the same Transformer LM both encodes the source and generates the summary. This ensures that all parameters in the network, including those governing attention over source states, have been pre-trained before the fine-tuning step. Experiments on the CNN/Daily Mail dataset show that our pre-trained Transformer LM substantially improves over pre-trained Transformer encoder-decoder networks in limited-data settings. For instance, it achieves 13.1 ROUGE-2 using only 1% of the training data (~3000 examples), while pre-trained encoder-decoder models score 2.3 ROUGE-2.

References in corpus (4)

• Sequence to Sequence Learning with Neural Networks
• On Using Monolingual Corpora in Neural Machine Translation
• Semi-supervised Sequence Learning
• Aligning Books and Movies: Towards Story-like Visual Explanations by Watching Movies and Reading Books

Cited by in corpus (15)

• PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization
• Leveraging Pre-trained Checkpoints for Sequence Generation Tasks
• Fine-Tuning Language Models from Human Preferences
• SAMSum Corpus: A Human-annotated Dialogue Dataset for Abstractive Summarization
• Big Bird: Transformers for Longer Sequences
• QURIOUS: Question Generation Pretraining for Text Generation
• Keyphrase Generation with Cross-Document Attention
• Product Title Generation for Conversational Systems using BERT
• Controlling the Amount of Verbatim Copying in Abstractive Summarization
• Multi-scale Transformer Language Models
• Generating Rich Product Descriptions for Conversational E-commerce Systems
• SciSummPip: An Unsupervised Scientific Paper Summarization Pipeline
• Understanding Neural Abstractive Summarization Models via Uncertainty
• Digging Deeper into CRNN Model in Chinese Text Images Recognition
• Exploiting News Article Structure for Automatic Corpus Generation of Entailment Datasets