第二阶段第二次会议

1. Category

1.1 Contrastive with Negative Samples

1) CPC [1]

2) CPC v2 [2]

3) CMC [3]

4) DIM [4]

5) DIM v2 [5]

6) PIRL [6]

7) MoCo [7]

8) SimCLR [8]

9) MoCo v2 [9]

10) InfoMin [10]

11) SimCLR v2 [11]

12) MoCo v3 [12]

1.2 Contrastive without Negative Samples

也有把这些工作称为“非对比学习”的说法

1) BYOL [13]

2) SwAV [14]

3) SimSiam [15]

4) Barlow Twins [16]

2. Disscusion: Potential Problems or Extension?

Scaling-up ability? – is stiil the bottleneck

Related work: SEER [17]

References

[1] Representation Learning with Contrastive Predictive Coding

[2] Data-Efficient Image Recognition with Contrastive Predictive Coding

[3] Contrastive Multiview Coding

[4] Learning deep representations by mutual information estimation and maximization

[5] Learning Representations by Maximizing Mutual Information Across Views

[6] Self-Supervised Learning of Pretext-Invariant Representations

[7] Momentum Contrast for Unsupervised Visual Representation Learning

[8] A Simple Framework for Contrastive Learning of Visual Representations

[9] Improved Baselines with Momentum Contrastive Learning

[10] What Makes for Good Views for Contrastive Learning

[11] Big Self-Supervised Models are Strong Semi-Supervised Learners

[12] An Empirical Study of Training Self-Supervised Vision Transformers

[13] Bootstrap Your Own Latent A New Approach to Self-Supervised Learning

[14] Unsupervised Learning of Visual Features by Contrasting Cluster Assignments

[15] Exploring Simple Siamese Representation Learning

[16] Barlow Twins: Self-Supervised Learning via Redundancy Reduction

[17] Self-supervised Pretraining of Visual Features in the Wild

视频

视频下载链接：https://pan.baidu.com/s/1A4A1LlZPzhotnvu_FUnBbg

提取码：hqaj

ppt

PPT下载链接：https://pan.baidu.com/s/1taeIKsUrywOLZg-zaK0K0g
提取码：t3jj