Conventional autoregressive left-to-right (L2R) sequence generation faces two issues during decoding: limited to unidirectional target sequence modeling, and constrained on strong local dependencies. To address the aforementioned problem, we propose P3LM, a probabilistically permuted prophet language model, which strengthens the modeling of bidirectional information and long token dependencies for sequence generation. Specifically, P3LM learns to generate tokens in permuted order upon an order-aware transformer decoder, as well as to generate the corresponding future N tokens with a multi-stream attention mechanism. Extensive experiments are conducted on the GLGE benchmark, which includes four datasets for summarization, two for question generation, one for conversational question answering, and one for dialog response generation, where P3LM achieves state-of-the-art results compared with strong publicly available generative pre-training methods.
@article{arxiv.2210.12339,
title = {P$^3$LM: Probabilistically Permuted Prophet Language Modeling for Generative Pre-Training},
author = {Junwei Bao and Yifan Wang and Jiangyong Ying and Yeyun Gong and Jing Zhao and Youzheng Wu and Xiaodong He},
journal= {arXiv preprint arXiv:2210.12339},
year = {2022}
}