Customs solves ethnic name identification headache

The Australian Customs Service is banking on a new identity checking system, which can distinguish between different permutations of foreign names, to help improve border security.Customs recently implemented a knowledge-based names database developed by Language Analysis Systems (LAS) to help its officers better identify individuals with names that do not translate naturally into the Roman alphabet.

The Australian Customs Service is banking on a new identity checking system, which can distinguish between different permutations of foreign names, to help improve border security.

Customs recently implemented a knowledge-based names database developed by Language Analysis Systems (LAS) to help its officers better identify individuals with names that do not translate naturally into the Roman alphabet. Such names can cause problems for law enforcement officials because most English-language databases assume that all names use Roman-characters and come in a first, middle and last name format, which is far from reality.

Peter Thompson, national manager of intelligence for the ACS, told ZDNet Australia  that names from different cultures can often be translated into several English equivalents. Because of the sheer volume of names and possible translations, regular databases find it virtually impossible to cope.

"For instance a simple Korean name Roh can quite legitimately be transliterated into English as Noh. So if you are looking for a Mr Noh from Korea and your electronic system has him down as Noh, but he has put Roh on his passport, then you are going to miss him," said Thompson.

Thompson said LAS' names database contains more than one billion names and is able to analyse a name and identify if it belongs to a particular gender, profession, religion, military status, culture or country. The system went live two weeks ago.

Jack Hermansen, chief executive of LAS, explained that because there are no dictionaries for names, problems arise when trying to identify names from different ethnic groups and geographic regions.

"With certain Arabic names, such as Abdul-Rahman or Abdul-Salaam, you don't want to match on the Abdul, or right side, because that is a prefix. It is the opposite when you look at Greek or Russian names where the ending is the same. With Romonov, Karpov, Andropov, the information is on the left side of the name," said Hermansen.

Australian Customs' Thompson said the LAS system will be used in conjunction with other technologies to improve the chances of identifying suspects.

"The system says that this name could also be spelt or transliterated into English in these forms -- you now have all the [names] and your chance of getting [someone on a watch list] will be much greater," said Thompson.

According to Hermansen, the LAS system is complementary to biometric systems such as fingerprinting and iris scanning because it is useful the first time you meet a person -- rather than the second.

"We make the point that biometrics are only useful the second time you meet somebody. The first time you don't have a retinal [or fingerprint] scan you say: 'ok we have your fingerprints, now behave yourself'. That is the best you can do," said Hermansen.

The LAS knowledge-based names database has already been used by the US Homeland Security Department's Customs and Border Protection agency and German bank WestLB AG, which uses the system to search for money launderers.

Newsletters

You have been successfully signed up. To sign up for more newsletters or to manage your account, visit the Newsletter Subscription Center.
See All
See All