TEncDM: Understanding the Properties of the Diffusion Model in the Space of Language Model Encodings
Alexander Shabalin, Viacheslav Meshchaninov, Egor Chimbulatov, Vladislav Lapikov, Roman Kim, Grigory Bartosh, Dmitry Molchanov, Sergey Markov
Abstract
This paper presents the Text Encoding Diffusion Model (TEncDM), a novel approach to diffusion modeling that operates in the space of pre-trained language model encodings. In contrast to traditionally used embeddings, encodings integrate contextual information. In our approach, we also employ a transformer-based decoder, specifically designed to incorporate context in the token prediction process. We conduct a comprehensive examination of the influence of the encoder, decoder, noise scheduler, and self-conditioning on zero-shot generation. Furthermore, we compare TEncDM with previous approaches on three conditional text generation tasks: QQP, XSum, and Wiki-Auto. The results show that TEncDM exhibits superior performance compared to existing non-autoregressive diffusion models.
BibTeX
@article{Shabalin_Meshchaninov_Chimbulatov_Lapikov_Kim_Bartosh_Molchanov_Markov_Vetrov_2025, title={TEncDM: Understanding the Properties of the Diffusion Model in the Space of Language Model Encodings}, volume={39}, url={https://ojs.aaai.org/index.php/AAAI/article/view/34696}, DOI={10.1609/aaai.v39i23.34696}, abstractNote={This paper presents the Text Encoding Diffusion Model (TEncDM), a novel approach to diffusion modeling that operates in the space of pre-trained language model encodings. In contrast to traditionally used embeddings, encodings integrate contextual information. In our approach, we also employ a transformer-based decoder, specifically designed to incorporate context in the token prediction process. We conduct a comprehensive examination of the influence of the encoder, decoder, noise scheduler, and self-conditioning on zero-shot generation. Furthermore, we compare TEncDM with previous approaches on three conditional text generation tasks: QQP, XSum, and Wiki-Auto. The results show that TEncDM exhibits superior performance compared to existing non-autoregressive diffusion models.}, number={23}, journal={Proceedings of the AAAI Conference on Artificial Intelligence}, author={Shabalin, Alexander and Meshchaninov, Viacheslav and Chimbulatov, Egor and Lapikov, Vladislav and Kim, Roman and Bartosh, Grigory and Molchanov, Dmitry and Markov, Sergey and Vetrov, Dmitry}, year={2025}, month={Apr.}, pages={25110-25118} }