Get the latest tech news
Learning from Heuristics
Summary I present a weak supervision paradigm called “data programming” which uses maximum likelihood estimation to produce soft labels from heuristics. These soft labels can then be used to train other models, without true labels being required at any stage.
I present a weak supervision paradigm called “data programming” which uses maximum likelihood estimation to produce soft labels from heuristics. I’ll use the BreastCancer dataset from the mlbench package since its well suited to binary classification and has a bunch of features to draw heuristics from. The results will be better or worse dependent on the labeling function coverage and precision, but I think the example demonstrates that the method can work well even with a simple baseline model.
Or read this on Hacker News