Tag: pseudo-labels

A new graph neural network framework called EGNN expands rare attack samples and uses threshold-controlled pseudo-labeling to improve intrusion detection ...

Researchers present CADS, a framework pairing a 22,022-scan dataset with complete annotations of 167 anatomical structures and a whole-body CT ...

Researchers report a neighbor contrastive learning method that fills missing data views by transferring neighbor relationships and improves clustering accuracy ...

A new method called Conflict-Aware Posterior Correction softens the hard masking used in partial-label learning, improving Top-1 accuracy by 1.01 ...

A semi-supervised framework that guides segmentation networks with signed distance maps achieved stronger boundary delineation on two breast cancer histopathology ...

A semi-supervised YOLO11 pipeline using confidence filtering, DIoU-NMS deduplication, and light manual checks outperforms fully supervised scrap steel detection while ...

A new semi-supervised AI framework called URDT-Net combines dual-teacher learning, topology-aware training, and an early-exit mechanism to segment abdominal organs ...

A new graph-structured Transformer model trained with self-supervised learning and community-consensus pseudo-labeling achieves record F1-scores for detecting cryptocurrency money laundering ...

Researchers in China have developed FullReg, a semi-supervised regression framework that weights pseudo-labels by data similarity and stabilizes training with ...

Researchers have developed a multi-granularity contrastive learning framework that automatically discovers new categories of teacher instructional behavior from classroom video, ...