See the signals – a route to using AI to spot leading indicators of environmental incidents: PAPER & PRESENTATION
Dr. Douglas Leith, COMET, UK
(free)
Organisations in high risk industries, such as water supply and wastewater management, gather large volumes of structured and unstructured data and information about incidents concerning their assets. This ranges from customer reports of issues, through internal observations and near miss reports, to full incident reports. Together this represents a considerable resource to assist in decision making
and understanding past, present and future risk.
Such organisations struggle to draw these sources together and extract meaningful intelligence. This can be due to lack of structure, or a common data structure; or systems unable to communicate together; or a lack of tools to analyse the data to extract actionable intelligence; or poor quality data and information.
Artificial intelligence (AI) tools, in particular natural language processing (NLP) and machine learning (ML), present an opportunity to make sense of this information allowing hypotheses to be tested, tracking trends, and pre-empting future incidents. The ultimate goal to achieve is therefore enhance decision making to better target investment and interventions. This paper outlines development using the COMET® AI module to ingest large data sources to learn and train models which turn data into actionable intelligence. This approach is illustrated with a case study in the wastewater industry.
