2022.3.7 论文速览

2022.3.7 论文速览
接下来的论文解读可能穿插更多的英文表述,懒,多多谅解。

文章目录

  • CAFE: Learning to Condense Dataset by Aligning Features(CVPR2022)
    • Method
    • Experiments
  • HCSC: Hierarchical Contrastive Selective Coding(CVPR2022)
    • Method
    • Experiments
  • BatchFormer: Learning to Explore Sample Relationships for Robust Representation Learning(CVPR2022)
    • Method

CAFE: Learning to Condense Dataset by Aligning Features(CVPR2022)

Method

本文主要目的是寻找一个替代数据集来减少模型的计算开销,具体做法是构建了一个合成数据集,构建方法如下:
2022.3.7 论文速览_第1张图片
Layer-wise feature alignment: minimize the difference of layer-wise feature maps of real and synthetic images using Mean Square Error (MSE).

To enable learning discriminative synthetic images, we use the feature centers of synthetic images of each class to classify the real images by computing their inner-product and cross-entropy loss.

Experiments

核心子集精度
2022.3.7 论文速览_第2张图片
可视化结果

HCSC: Hierarchical Contrastive Selective Coding(CVPR2022)

Method

现有的对比学习方法没有考虑到图片的分层特性
2022.3.7 论文速览_第3张图片

Experiments

框架结构:有一说一,这个分层的先验知识要求会不会太强了
2022.3.7 论文速览_第4张图片
部分实验结果:
2022.3.7 论文速览_第5张图片

BatchFormer: Learning to Explore Sample Relationships for Robust Representation Learning(CVPR2022)

Method

2022.3.7 论文速览_第6张图片

你可能感兴趣的:(论文分享,计算机视觉,深度学习,人工智能)