Text Diffusion with Reinforced Conditioning
This work addresses limitations in text diffusion models for non-autoregressive sequence generation, offering incremental improvements for natural language processing tasks.
The paper tackled the problem of text diffusion models underperforming due to issues with self-conditioning degradation and training-sampling misalignment, proposing TREC with Reinforced Conditioning and Time-Aware Variance Scaling, which demonstrated competitive performance against various baselines in experiments.
Diffusion models have demonstrated exceptional capability in generating high-quality images, videos, and audio. Due to their adaptiveness in iterative refinement, they provide a strong potential for achieving better non-autoregressive sequence generation. However, existing text diffusion models still fall short in their performance due to a challenge in handling the discreteness of language. This paper thoroughly analyzes text diffusion models and uncovers two significant limitations: degradation of self-conditioning during training and misalignment between training and sampling. Motivated by our findings, we propose a novel Text Diffusion model called TREC, which mitigates the degradation with Reinforced Conditioning and the misalignment by Time-Aware Variance Scaling. Our extensive experiments demonstrate the competitiveness of TREC against autoregressive, non-autoregressive, and diffusion baselines. Moreover, qualitative analysis shows its advanced ability to fully utilize the diffusion process in refining samples.