RegGenome, a University of Cambridge commercial spin-out, is a regulatory data technology company and a leader in the field of computational regulation. We are changing how the world produces and consumes regulatory information. As a regulatory data provider, we process the world’s regulation using AI to transform what is human-readable into machine-readable and machine-consumable data.
Mark Johnston, CEO #judge #grandfinale
Mark Gemmell – Head of Engineering #judge #firstround
Emmanuel (Manos) Schizas, Head of Regulatory Engineering TECH WORKSHOP: ‣
Shruti Kumar, Director of Commercial Operations #mentor
Victor Chuah, Head of Data Crawling
Max Ashton-Lelliott, Senior Data Scientist
Toby Cassidy, Senior Data Scientist
The largest regulated financial services firms and their service providers sift through tens of thousands of regulatory notifications annually, at significant cost.
Each group knows that the majority of these notifications are bound to be irrelevant to most of their constituents’ operations but are also wary of missing relevant content. Many therefore use a (largely manual) triaging system to score the relevance of each document to business lines, functions and products. The result is a great deal of redundant work that ties down significant compliance generalist and occasionally specialist resources without always achieving commensurate reductions in risk.
Unfortunately, the granular elements of legal and regulatory text – clauses, sections, etc - typically do not name the types of entities to which each obligation applies, confounding simple NLP based classification. Some publishers are helpful enough to produce glossaries of defined terms within the texts themselves but in many cases the terms and their definitions are ultimately set out in upstream legislation or key rules and must be inferred onto downstream documents, including detailed rules or guidance. Hierarchies of products, such as ISO10962 (see also here), the Common Domain Model (CDM), the ISDA taxonomy or the Financial Instrument Global Identifier (FIGI) are very useful analytical tools but are designed and indeed largely reserved for manual application.
Finally, the perimeter of financial regulations is in many cases controversial or hard to define and clarifying the scope of regulated activities or the reach of regulators’ powers is frequently the object of legal advice.
Participants should demonstrate an efficient, technology-enabled method for recognising what types of firms, products, services, or activities each incoming rule or legal provision applies to and organise these into a simple hierarchical structure.
Participants will be provided with a set of RegGenome structured documents, complete with document and sub-document level metadata, as well as links to their origin URLs. They will be asked to derive from this dataset: