ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training

13 Jan 2020  ·  Weizhen Qi, Yu Yan, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang, Ming Zhou ·

This paper presents a new sequence-to-sequence pre-training model called ProphetNet, which introduces a novel self-supervised objective named future n-gram prediction and the proposed n-stream self-attention mechanism. Instead of optimizing one-step-ahead prediction in the traditional sequence-to-sequence model, the ProphetNet is optimized by n-step ahead prediction that predicts the next n tokens simultaneously based on previous context tokens at each time step. The future n-gram prediction explicitly encourages the model to plan for the future tokens and prevent overfitting on strong local correlations. We pre-train ProphetNet using a base scale dataset (16GB) and a large-scale dataset (160GB), respectively. Then we conduct experiments on CNN/DailyMail, Gigaword, and SQuAD 1.1 benchmarks for abstractive summarization and question generation tasks. Experimental results show that ProphetNet achieves new state-of-the-art results on all these datasets compared to the models using the same scale pre-training corpus.

PDF Abstract

Results from the Paper

Ranked #6 on Question Generation on SQuAD1.1 (using extra training data)

     Get a GitHub badge
Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Result Benchmark
Abstractive Text Summarization CNN / Daily Mail ProphetNet ROUGE-1 44.20 # 17
ROUGE-2 21.17 # 18
ROUGE-L 41.30 # 13
Text Summarization GigaWord ProphetNet ROUGE-1 39.51 # 7
ROUGE-2 20.42 # 7
ROUGE-L 36.69 # 8
Question Generation SQuAD1.1 ProphetNet BLEU-4 23.91 # 6
METEOR 26.6 # 3
ROUGE-L 52.3 # 4