Injecting Logical Constraints into Neural Networks via Straight-Through Estimators

Open in new window