Artificial intelligence improves Norwegian Customs data base

What are we doing?

We are working to ensure that the declaration on goods that are on their way into and out of Norway are correct.

This is an artificial intelligence (AI) project where we carry out machine learning on customs declarations. With the help of AI, we compare sender and recipient with previous goods declarations for the same ones, and find similarities between them.

Norwegian Customs meet the daily task of receiving and managing declarations for goods crossing the Norwegian border. However, the manual completion of these declarations increases the likelihood of errors due to potential typos or inconsistencies from different sources. Consequently, maintaining an accurate and up-to-date record of foreign entities and associated declarations poses a significant challenge. The sheer volume and frequency of incoming declarations, combined with the extensive storage requirements, make manual validation or basic fuzzy matching impractical, and this is where artificial intelligence comes into play.

What is entity resolution?

Entity resolution is a branch of artificial intelligence that aims to identify data records that belong to the same real object. The applications span various areas, including matching of customer registers, health records, financial transactions and, in this context, identification of foreign senders and receivers. This is especially valuable when merging data from different sources, helping with customer management and fraud detection. Norwegian Customs benefit from entity resolution by detecting similar foreign units, thereby capturing a more comprehensive data representation of the flow of goods.

Why do we do this for the Norwegian Customs Service?

This is done to ensure that the product declarations that are registered are correct. For the Norwegian Customs Agency, which annually declares large quantities of goods in and out of the country, this is of great importance. Our machine learning algorithms try to correct the inaccuracies that occur when product declarations are filled in and read. The algorithms correct the inaccuracies and or errors found in already completed declarations.

This applies to all types of declarations from both private and professional purchases and/or sales.