顶级会议论坛PPT：陈挺.pptx

《顶级会议论坛PPT：陈挺.pptx》由会员分享，可在线阅读，更多相关《顶级会议论坛PPT：陈挺.pptx（38页珍藏版）》请在三个皮匠报告上搜索。

1、Ting Chen Simon Kornblith Mohammad Norouzi Geoffrey Hinton,SimCLR: A Simple Framework for Contrastive Learning of Visual Representations,Google Research, Brain Team,Unsupervised representation learning,We tackle the problem of general visual representation learning from a set of unlabeled images. Af

2、ter unsupervised learning, the learned model and image representations can be used for downstream applications,Unlabeled data (images,Unsupervised pretrained network,Downstream applications,First category of unsupervised learning,Generative modeling Generate or otherwise model pixels in the input sp

3、ace Pixel-level generation is computationally expensive Generating images of high-fidelity may not be necessary for representation learning,Image credit: Xifeng Guo, Thalles Silva,Autoencoder,Generative Adversarial Nets,Second category of unsupervised learning,Discriminative modeling Train networks to perform pretext tasks where both the inputs and labels are derived from an unlabeled dataset. Heu