As part of its ongoing efforts to improve nuclear safety, IRSN decided to call on Novelis to assist it in setting up a feedback processing platform with an analysis enhanced by Artificial Intelligence.
The client's challenges and objectives
IRSN, the French Institute for Radiation Protection and Nuclear Safety, conducts research and expert assessments on the risks associated with radioactivity.
In order to process significant events reported by French nuclear operators, IRSN has launched the PIREX project: Integrated Feedback Platform.
This project consists of setting up this platform to:
Decompartmentalize the processing of feedback
Drawing lessons to effectively guide IRSN’s expert assessment work
Using artificial intelligence, propose thematic analyses useful to all stakeholders involved in the continuous improvement of nuclear safety: administrative authorities, technical experts, operators, civil society
We have supported the IRSN in its PIREX project with our partner ABBYY and its FlexiCapture tool for intelligent document processing.
The artificial intelligence engine of the PIREX platform is fed through OCR and IDP (Intelligent Document Processing) extraction of data from the operators’ reports. The data in these reports is unstructured data.
Thanks to these technologies, we have assisted the IRSN in the implementation of an end-to-end automation of the analysis process of the nuclear centers’ operating reports.
Novelis' added value
Our expertise in Smart Automation
Through our expertise in AI, NLP and OCR, we are able to help our clients to deliver efficient automation projects: Smart Automation.
This means taking into account a general vision of the technology and its contributions in order to deliver a relevant solution to manage both structured and unstructured data, always with the concern of being aligned with the IT roadmap and impacting the economic performance (ROI).
IT project manager - IRSN
Novelissupported IRSN in an efficient and professionalmanner in the deployment of a Flexicapture solution and the definition of recognition models for long and complex documents.