Container shipping blockchain start-up Blockshipping has partnered with enterprise AI company 2021.AI on the creation of a new product using artificial intelligence (AI) to predict import dwell times.
Blockshipping says that the AI-based Import Dwell Time Prediction (IDP) model will be the first in a series of AI models that aims to provide industry stakeholders with AI-based end-to-end supply chain visibility, focusing specifically on improving container ETAs.
The new product has been built leveraging 2021.AI’s Grace platform and data science expertise, married with Blockshipping’s experience in the container shipping market.
“The partnership has been very successful because of Blockshipping’s clear commitment to the business problems at hand,” said Rasmus Hauch, Head of Product Management at 2021.AI.
“We found a great match in combining the business perspective with our experience in taking AI into production on our AI Platform, Grace. This means that we are not only developing a single model but building an interface for future scaling of AI models for terminal operators. Our partnership with Blockshipping is a great example of how AI can bring value by solving real business challenges.”
The IDP service aims to forecast how long containers will remain in a container terminal before being delivered to the importer, based on analysis of large amounts of historical data collected from carriers and terminals.
The data is analysed in collaboration with the terminal by Blockshipping’s internal domain experts, weighing the parameters that are estimated to have a significant impact on container dwell time and using that information to train the AI model.
The company says that the system will operate similar to a neural network, gradually improving through training to create better and better mathematical models. IDP will initially allow the trained neural network to be queried about dwell time predictions via a web browser, though planned future versions are planned to provide direct integration to the carrier’s container management system and the terminal’s TOS.
Blockshipping says that IDP will also use traditional statistical methods, continuously updated with new data to progressively learn from the deviations that might occur between the model’s predictions and actual dwell times. As a result, the predictions should continually improve and become more precise.
The Danish company notes that it has been granted new funding to support the development of IDP by The Danish Maritime Fund.