ACM CIKM · 2026 · LACE
LACE: Unsupervised Concept Drift Detection in Multi-Label Data Streams Through Label Cluster Evolution
Gofralilar, M. K., Ghahramanian, P., & Can, F.
LACE detects concept drift in multi-label data streams without access to labels, by tracking how clusters of co-occurring labels evolve over time.
Against the previous best unsupervised method it reduces detection delay by 63.5% while maintaining a 0% missed-detection rate.