CargO2ai for the Port of Montréal
Partner
Context
With 38 million tons of cargo arriving annually, the Port of Montreal is the second largest port in Canada and the fifth largest container port on the Eastern coast of North America, serving Quebec, Ontario and the United States. The Port of Montreal deals with an abundance of data on a daily basis and the organization took it as an opportunity to connect and leverage its important data to make better decisions. Over the past few years, the Port of Montreal benefited from its digital maturity and has worked hard to position itself as one of the smartest ports in the world.
Project overview
CargO2ai was launched in the spring of 2020, right after COVID-19 pandemic hit. This AI initiative was developed to identify and prioritize marine cargo before vessels arrive in port, so that life-saving protective equipment and essential medical supplies were redirected as quickly as possible to hospitals and other facilities.
The challenge
In the context of COVID-19, the CargO2ai initiative was Port of Montreal’s response to the life-or-death issue of prioritizing the identification and shipping of PPE, medications and medical equipment arriving at the Terminals. Before arriving at the Port, each vessel sends a cargo manifest including a description of all the items to be found on board. Each of those cargo manifests includes millions of lines of data, which makes it difficult to identify and select specific items. In a situation where specific equipment is urgently needed–say in the context of COVID-19–the lack of a quick identification process creates ad hoc processes and potential delays for the end users–in this case, hospitals and patients.
Oops
You need to accept marketing COOKIES to WATCH THIS video.
The solution
That’s where the power of AI comes in. Port of Montreal and its partners CargoM, Termont, MGTP, Scale AI and Ivado Labs worked on implementing an algorithm that uses Natural Language Processing to recognize text and look for keywords and numerical codes in the extensive unstructured data of the cargo manifest. The algorithm can understand specific keywords within the description’s context to detect which containers should be prioritized more accurately. In the context of COVID, this allowed Port of Montreal employees to detect the location of medical equipment in an instant, and to make it available for quick shipping to hospitals. Given the situation’s urgency, the project was launched and completed in a record-breaking 12 weeks.
Stats
Results and Benefits
A solution that is both modular and scalable
Human validation is still required to ensure that the right material is moved, but the algorithm learns and continuously improves. It’s also possible to add more keywords, and expand the use for other merchandise, which means that it could be applied to other materials, industries and critical situations.
An impactful project
The solution had the positive side effect of improving communication between the Port of Montreal and its global partners, giving access to data way sooner in the process, which significantly increases predictability and efficiency.
"One of the great challenges is that there is an overabundance of data and that we need very short response times. We need algorithms to quickly process information, extract what is important and manage to support decision-making."
Scientific Director at IVADO Labs
Technologies
The Partners
This project was realized in partnership with Port of Montréal, CargoM, Termont, MGTP, Scale AI.
