Doesn't feel fair to compare a model built on less data given sampling is informed by first training on the full data...
New tutorial on http://keras.io : active learning, an effective way to optimize your investment in data annotation. https://keras.io/examples/nlp/active_learning_review_classification/ …
Created by @getdarshanpic.twitter.com/T4JmEVEo4u
-
-
-
How is the sampling of data points informed by training on the full data? Active Learning (AL) is a greedy technique. The model trained on the full 40,000 reviews is just for comparison and has no connection with the training process of the AL model whatsoever.
- Show replies
New conversation -
-
-
How to train custom dataset for adverse drug event in keras using GNN
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.