Improved Paraphrase Generation via Controllable Latent Diffusion
This work addresses a specific bottleneck in text generation for NLP applications, offering incremental improvements in paraphrase generation.
The paper tackles the truncation issue in textual diffusion models for paraphrase generation by proposing LDP, a controllable latent diffusion method that improves efficiency and semantics control, achieving superior quality and diversity compared to baselines.
Paraphrase generation strives to generate high-quality and diverse expressions of a given text, a domain where diffusion models excel. Though SOTA diffusion generation reconciles generation quality and diversity, textual diffusion suffers from a truncation issue that hinders efficiency and quality control. In this work, we propose \textit{L}atent \textit{D}iffusion \textit{P}araphraser~(LDP), a novel paraphrase generation by modeling a controllable diffusion process given a learned latent space. LDP achieves superior generation efficiency compared to its diffusion counterparts. It can facilitate only input segments to ensure paraphrase semantics, improving the results without external features. Experiments show that LDP better reconciles paraphrase generation quality and diversity than baselines. Further analysis shows that our method is also helpful to other similar text generations and domain adaptations