English

InsNet: An Efficient, Flexible, and Performant Insertion-based Text Generation Model

Computation and Language 2022-10-18 v4

Abstract

We propose InsNet, an expressive insertion-based text generator with efficient training and flexible decoding (parallel or sequential). Unlike most existing insertion-based text generation works that require re-encoding of the context after each insertion operation and thus are inefficient to train, InsNet only requires one pass of context encoding for the entire sequence during training by introducing a novel insertion-oriented position encoding and a light-weighted slot representation strategy to enable computation sharing. Furthermore, we propose an algorithm InsNet-Dinic to better determine the parallelization of insertion operations that provides a controllable switch between parallel and sequential decoding, making it flexible to handle more parallelizable tasks such as machine translation with efficient decoding, or less parallelizable tasks such as open-domain text generation to guarantee high-quality outputs. Experiments on two lexically constrained text generation datasets and three machine translation datasets demonstrate InsNet's advantages over previous insertion-based methods in terms of training speed, inference efficiency, and generation quality.

Keywords

Cite

@article{arxiv.2102.11008,
  title  = {InsNet: An Efficient, Flexible, and Performant Insertion-based Text Generation Model},
  author = {Sidi Lu and Tao Meng and Nanyun Peng},
  journal= {arXiv preprint arXiv:2102.11008},
  year   = {2022}
}

Comments

Accepted as a poster paper at NeurIPS 2022

R2 v1 2026-06-23T23:24:00.685Z