References & Citations
Computer Science > Computation and Language
Title: Continuous diffusion for categorical data
(Submitted on 28 Nov 2022 (v1), last revised 15 Dec 2022 (this version, v3))
Abstract: Diffusion models have quickly become the go-to paradigm for generative modelling of perceptual signals (such as images and sound) through iterative refinement. Their success hinges on the fact that the underlying physical phenomena are continuous. For inherently discrete and categorical data such as language, various diffusion-inspired alternatives have been proposed. However, the continuous nature of diffusion models conveys many benefits, and in this work we endeavour to preserve it. We propose CDCD, a framework for modelling categorical data with diffusion models that are continuous both in time and input space. We demonstrate its efficacy on several language modelling tasks.
Submission history
From: Sander Dieleman [view email][v1] Mon, 28 Nov 2022 06:08:54 GMT (350kb,D)
[v2] Tue, 6 Dec 2022 15:52:02 GMT (350kb,D)
[v3] Thu, 15 Dec 2022 14:27:19 GMT (350kb,D)
Link back to: arXiv, form interface, contact.