Keep the Needle, Prune the Haystack: Defect-Preserving Token Pruning for Efficient Zero-Shot Anomaly Detection 文章

ArXiv CS.CV2026-08-05PAPERen作者: Yanning Hou, Jingyuan Zhang, Xiaoyun Wang, Qixiang Ma, Sihang Zhou, Ke Xu
查看原文 →
关系图谱

Keep the Needle, Prune the Haystack: Defect-Preserving Token Pruning for Efficient Zero-Shot Anomaly Detection · 相关技术

相关技术

AWAReCuPyConfidence-Guided Early StoppingCalibrated Entropy ScoreCTICRISCOMBayesian Membership Inference AttackBaseline (zero-shot)Auxiliary Regression LossAutomatic Text ScoringAugmented Latent IntrinsicsAudio Mamba EncoderArtificial Language LearningArabic Speech Emotion RecognitionAmodal Mask Prediction and ROI Re-Alignment (AMPR)Adaptive Batch ScalingATEARTARIANOANNAHATABLE3D Segment Anything Model3D LUTAbstractACT