A bimodal framework for nonstationary process monitoring via collaborative contrastive and adversarial unsupervised learning.

Recognizing nonstationarity is pivotal for trustworthy industrial process monitoring. Existing methods address this issue from a unimodal perspective, which struggles to capture intrinsic heterogeneity. To resolve this, we introduce a novel unsupervi…

Preview only. Read the full abstract at the source