$\textit{latent}$-GLAT: Glancing at Latent Variables for Parallel Text Generation

TL;DR

This paper introduces $ extit{latent}$-GLAT, a parallel text generation method using discrete latent variables and curriculum learning, which improves generation quality without relying on autoregressive models, expanding application possibilities.

Contribution

It proposes a novel parallel text generation approach with discrete latent variables and curriculum learning, eliminating the need for autoregressive training.

Findings

Outperforms strong baselines in quality

Does not require autoregressive models for training

Broadens application scenarios of parallel decoding

Abstract

Recently, parallel text generation has received widespread attention due to its success in generation efficiency. Although many advanced techniques are proposed to improve its generation quality, they still need the help of an autoregressive model for training to overcome the one-to-many multi-modal phenomenon in the dataset, limiting their applications. In this paper, we propose $\textit{latent}$-GLAT, which employs the discrete latent variables to capture word categorical information and invoke an advanced curriculum learning technique, alleviating the multi-modality problem. Experiment results show that our method outperforms strong baselines without the help of an autoregressive model, which further broadens the application scenarios of the parallel decoding paradigm.