Tweeted By @ml_review
Contrastive Multiview Coding
— ML Review (@ml_review) June 17, 2019
By @YonglongT @dilipkay @phillip_isola
Unsupervised representation learned to maximize MI measure between different views results in SoTA for downstream tasks such as object classification
ArXivhttps://t.co/xlk1C7l1FZ
Githubhttps://t.co/S1vpmgjuIR pic.twitter.com/uCMuyOQJrX