These plots (also included in the updated version of our UDA paper https://arxiv.org/abs/1904.12848 with a lot more results & details) illustrate very well Vincent's article on the quiet revolution of semi-supervised learning!https://towardsdatascience.com/the-quiet-semi-supervised-revolution-edec1e9ad8c …
-
-
Prikaži ovu nit
-
Forgot to mention: left plot is IMDb sentiment analysis (accuracy); right plot is CIFAR10 semi-supervised learning benchmark (error rate). All these results can be reproduced with the code in https://github.com/google-research/uda … and read more athttps://ai.googleblog.com/2019/07/advancing-semi-supervised-learning-with.html …
Prikaži ovu nit
Kraj razgovora
Novi razgovor -
-
-
Great work! While the framework of UDA is general, the techniques introduced (back-translation, word replacement) focuses on text classification. What about token level and span level tasks?
-
We have been actively exploring Cross View Training for NER https://arxiv.org/abs/1809.08370 and would love to know other alternatives!
Kraj razgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.