A curriculum-based test-time adaptation method, using balanced confident pseudo-labels and multi-scale refinement, improves AIGC detector accuracy on unseen generators by 11 to 29 points over its starting detector.
Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Test-Time Curriculum for Open-Set AIGC Detection
A curriculum-based test-time adaptation method, using balanced confident pseudo-labels and multi-scale refinement, improves AIGC detector accuracy on unseen generators by 11 to 29 points over its starting detector.