成學

This is a review of "Distilling Knowledge via Knowledge Review" presented at CVPR 2021.

This is a review of"Decoupled Knowledge Distillation"presented at CVPR 2022.IntroductionTypes of Knowledge DistillationLogits-based method$(+)$ Computational and storage cost ↓$(-)$ Unsatisfactory performanceFeature-based method$(+)$ Superior performance$(-)$ Extra computational cost and storage usage∴ Potential of logit distillation is limited. Decoupled Knowledge DistillationTarget classifiati..

This is a review of "Similarity-Preserving Knowledge Distillation" presented at ICCV 2019.

This is a review of "Distilling the Knowledge in a Neural Network" presented at arXiv 2015.

This is a review of "RobustNet: Improving Domain Generalization in Urban-Scene Segmentation via Instance Selective Whitening" presented at CVPR 2021.

보호되어 있는 글입니다.

This content, created around August 2022, is a summary of Stanford University's CS231n course.