How we do it
We approach regulation as content, not as code. Because regulation as code is difficult to execute outside of narrow use-cases, since it implies regulatory language is unambiguous. We are sequencing the world’s regulatory information and will make this openly accessible as an innovation platform on which interoperable software applications can be built and that will foster knowledge sharing and creation.
Build a repository of regulatory data.
Domain experts collaborate to build a taxonomy of relevant information covered in the documents.
3. Annotated training dataset
Human experts annotate key information in documents according to the taxonomy, ultimately building an annotated training dataset.
4. Classifier model
Annotated documents are used to train a classifier model using AI/ML.
Trained classifiers machine-sequence regulatory documents, with sequenced data output to an API.
Interoperable applications access the machine readable data through the API.
Machine-readable regulation: an introduction
Emmanuel Schizas, Head of Product at Regulatory Genome Project and Lead in Regulation and RegTech at Cambridge Centre for Alternative Finance, Cambridge Judge Business School, offering an intro on our approach to machine-readable regulations.
Join our team
If you enjoy working in a dynamic, collaborative environment where new ideas and creativity are always encouraged, view our open positions.