2019-06-03, 10:30–10:55, Hall 2 (PyData)
Most people would not consider the language written within legal documents to be "natural" to any human being. We would demonstrate how their structure can be proved handy with several NLP techniques.
Legal documents are known to be obscure, repititive and overly-complicated.
In this talk we would cover our work at BestPractix, and demonstrate how the structure of legal contracts can be used for various tasks.
We would cover: (0) Data annotation hacks (1) fast clause and contract classification (2) Identifying defined terms (3) Applying deep learning to learn style.