Variational diffusion representation learning framework for semi-supervised industrial soft sensing application

Hengqian Wang

doi:10.1016/j.jmsy.2025.11.013

Variational diffusion representation learning framework for semi-supervised industrial soft sensing application

2025-11-18

0

PRE

AI

DOI：10.1016/j.jmsy.2025.11.013

Original

Original request for help

Save

Abstract

En 中文

• The first end-to-end diffusion model-based representation framework for semi-supervised soft-sensor. • Parametric terminal variational diffusion model is proposed for general temporal feature extraction. • Supervised temporal-diffusion joint representation model is proposed for quality prediction. • The effectiveness and stability of VDRL framework are validated in SRU and polymerization processes.

Variational diffusion representation learning framework for semi-supervised industrial soft sensing application

Abstract

Journal

Organization