Information Extraction with Neural Networks and Free Noisy Supervision

How does the pipeline compare to state-of-the-art information extraction suites (e.g. Apache OpenNLP) on standard datasets?

Where's the code?

Well. That was a short paper.