Direct naar content

You are here:

  1. Home
  2. Algorithms
  3. Act AI

Please note: The algorithm descriptions in English have been automatically translated. Errors may have been introduced in this process. For the original descriptions, go to the Dutch version of the Algorithm Register.

Act AI

Land Registry

Deed AI ensures that data from notarial deeds is automatically transferred. The employee processing the deed only has to check the data and adjust it if necessary.

Last change on 4th of June 2024, at 11:20 (CET) | Publication Standard 1.0
Publication category
Other algorithms
Impact assessment
Field not filled in.
Status
In use

General information

Theme

  • Living
  • Space and Infrastructure

Begin date

Field not filled in.

Contact information

https://formulieren.kadaster.nl/contact_kadaster_algoritmeregister

Link to source registration

https://www.kadaster.nl/over-ons/beleid/algoritmeregister/akte-ai

Responsible use

Goal and impact

With Deed AI, we want to be able to process deeds as quickly and correctly as possible. It helps us work efficiently and improve the quality of deed processing.

Considerations

Act AI has no negative impact on society. The process is designed so that legal certainty is not diminished and probably even improved.

Human intervention

Human in the loop. The Act AI makes a suggestion to the employee processing the data. So there is always an employee, who evaluates the suggestion.

Risk management

Minimal risk

Legal basis

Under Section 3(1a) of the Land Registry Act, it is our statutory duty to manage and update the Basic Land Registry. That is where Act AI helps.

Operations

Data

For recognising personal data, we use "bert-base-dutch-cased-finetuned-sonar-ner". This is an open source language model from the University of Groningen.For recognising other data, we use models we have trained ourselves. And algorithms we have developed ourselves.For recognising personal data, we use "bert-base-dutch-cased-finetuned-sonar-ner". This is an open source language model from the University of Groningen.For recognising the other data, we use models we have trained ourselves. And algorithms we have developed ourselves. About the training sets The training sets consist of actuals. These are currently still static and vary in size: from a few hundred to several tens of thousands of texts per set. Furthermore, we are working on a feedback loop. With this, it will no longer be a static set: every month, 8.33% will be refreshed. Texts will disappear and new action texts will replace them.

Technical design

  • Named Entity Recognition of personal names with Hugging Face Transformers
  • Named Entity Recognition of organisation name, address and legal form with the spaCy framework
  • Dependency Parsing of persons, organisations, real estate, notaries, legal facts and signature with the spaCy framework
  • Named Entity Recognition of property description with the spaCy framework
  • Named Entity Recognition for splitting personal names into first name, middle name and surname with the spaCy framework
  • Named Entity Recognition for splitting an address into street, house number, house letter, and addition with the spaCy framework
  • Classification of deed type with the spaCy framework
  • The self-developed algorithms use, among others:
    • regular expressions
    • spaCy pattern matchersNamed Entity Recognition of personal names with Hugging Face Transformers

Similar algorithm descriptions

  • Information extraction from, and classification of, notarial deeds (DNI)

    Dutch Tax and Customs Administration

    The algorithm detects and selects text from notarial deeds for monitoring the correct registration of the notarial deed and levying and collecting taxes.

    Last change on 25th of June 2024, at 10:22 (CET) | Publication Standard 1.0
    Publication category
    Impactful algorithms
    Impact assessment
    DPIA
    Status
    In use

  • Enter notifications

    Regional Environment Agency North Sea Canal Area

    With the deployment of this automation, notifications are automatically put into the case system where a case is created.

    Last change on 27th of February 2024, at 9:49 (CET) | Publication Standard 1.0
    Publication category
    Impactful algorithms
    Impact assessment
    Field not filled in.
    Status
    Out of use

  • Anonymisation software

    Province of Zeeland

    The algorithm underlines personal data in documents. An employee has to look at all the pages and check whether the document is properly anonymised. Then the software removes all underlined information and varnishes it. After that, the documents can be published, for example under the Open Government Act.

    Last change on 14th of January 2025, at 15:11 (CET) | Publication Standard 1.0
    Publication category
    Impactful algorithms
    Impact assessment
    Field not filled in.
    Status
    In use

  • Anonymisation software

    Municipality of Beekdaelen

    The algorithm underlines personal data in documents. An employee has to review all the pages and check whether the document is properly anonymised. Then the software removes all marked information and blacklists it irretrievably. After that, the documents can be published, e.g. under the Open Government Act.

    Last change on 30th of September 2024, at 13:43 (CET) | Publication Standard 1.0
    Publication category
    Other algorithms
    Impact assessment
    DPIA, ...
    Status
    In use

  • Anonymisation software

    Municipality of Schagen

    The algorithm underlines personal data in documents. An employee has to review all the pages and check whether the document is properly anonymised. Then the software removes all highlighted information and blacklists it. After that, the documents can be published, for example under the Open Government Act.

    Last change on 19th of September 2024, at 9:10 (CET) | Publication Standard 1.0
    Publication category
    Other algorithms
    Impact assessment
    DPIA, ...
    Status
    In use