Intelligent Analysis of Regulatory Documents using Deep Neural Networks

The Environment, Health, and Safety (EHS) documents should be deconstructed by experts to create action items to track compliance performance, audits, or for compliance certification. Modifications in the regulations, changes in the site conditions, mergers and acquisitions occur frequently and should be processed by experts. The manual processing of the text documents to extract different components of a regulatory content including tables, cover pages, etc requires a lot of effort, and is time consuming and error prone. The automation of this process helps reducing these drawbacks. The automatic extraction of these components is challenging due to non-standardized text structure and heterogeneous documents. This research project will provide solutions for automatic extraction of different components from EHS documents using machine learning and natural language processing techniques. The results of this project are new algorithms and automation tools and will benefit the ehsAI and Canadian companies in the EHS fields directly. The developed techniques are expected to be applied in other areas such as Software companies.

Faculty Supervisor:

Luc Lamontagne

Student:

Partner:

ehsAI

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

Université Laval

Program: