In late November, a consortium consisting of Krakow group Comarch Poland and Wroclaw Neurosoft signed an agreement to implement a common court system for the publication of rulings.
As representatives of the Wroclaw company, a participant in Polish Champion, explain, advanced techniques for the analysis of grammatical texts allow the implemented system to significantly accelerate the process of so-called anonymization (the removal of selected information, e.g. personal data), which will allow almost immediate publication of rulings by courts throughout the country.
The project is financed by the Ministry of Justice and concerns the implementation of the system in the 182 courts in which such a solution is not yet operational. As a result, by the end of January 2015, the system will cover all courts in Poland.
Currently, this type of tool works in more than 120 courts and contains the full content of more than 80 thousand rulings publicly available on the internet in one place, in the so-called judgment portal.
For over a year, this knowledge base has been used by more and more people – every month the portal is visited by over 270 thousand users, generating over 1 million views.
Neurosoft’s representatives explain that access by the general public to court undoubtedly has a significant impact on the quality of law.
“Less than two years ago it was still difficult to publicize current rulings, because it was necessary to remove manually sensitive data from the text – this process was not just time consuming but also burdened with a high risk of errors. In addition, manual anonymization result was often inconsistent” – they comment.
Thanks to the cooperation of the scientific community and Neurosoft, it has been possible to create a system for the automatic removal of sensitive content, and the human role is limited to verification of computer-implemented changes.
“The construction of such an advanced system would not have been possible without close cooperation between business and science” emphasizes Cezary Dołęga, R&D Director in Neurosoft.
With the support of researchers in the University of Wroclaw, it was first possible to analyze the range of anonymization practices used by the courts that created an extensive set of anonymization rules.
Then these rules were “translated” into the computer and the result is a specialized program that automatically dispenses the rules for any text of the ruling.
Anonymization is not the only unique feature of the developed system. The program also finds references to legislation, invents thematic entries in the analyzed texts and can automatically find similar rulings – a ruling related to a specified standard.