Papers2

#confidence thresholding

Toward Stable Semi-Supervised Remote Sensing Segmentation via Co-Guidance and Co-Fusion

Co2S is a new way to train segmentation models with very few labels by letting two different students (CLIP and DINOv3) learn together and correct each other.

#semi-supervised segmentation#remote sensing#pseudo-label drift

Learning Unmasking Policies for Diffusion Language Models

Intermediate

Metod Jazbec, Theo X. Olausson et al.Dec 9arXiv

Diffusion language models write by gradually unmasking hidden words, so deciding which blanks to reveal next is a big deal for both speed and accuracy.

#diffusion language models#masked diffusion#unmasking policy