Synch, the law firm behind Legal AI Blog, is once again committed to hosting Global Legal Hackathon in Stockholm.
As a modern law firm with technology at its core, and as the first law firm in the Nordics to have developed a client-oriented legal AI app entirely in-house, Synch and its technology partners will make available to the participants a developer toolbox.
The developer toolbox will include free API access to blockchain transactions and data extraction library.
Blockchain API from iDefendo
One of Synch technology partners iDefendo, a blockchain company providing reliable evidence to help intellectual property-intensive businesses prove ownership of their content, will provide the participants with free API access to blockchain transactions (“digital witnesses”).
Endpoints that will be available to the participants will allow to create new digital witnesses, chain digital witnesses and retrieve the existing ones.
iDefendo will provide example scripts in Postman that can be used as templates in projects and to get acquainted with the system.
Digital Witnesses API will be helpful to the teams who intend to use blockchain in their projects, allowing the participants to build a proof-of-concept faster.
Synch LegalNLP is a library for marking (extracting/highlighting) key information in a legal document. The following categories of information are marked:
- amounts (monetary amounts, periods, other units, raw amounts, percentages)
- phone numbers
- governing law
Additional functionality provided by Synch LegalNLP includes:
- strip HTML from text – submit an HTML string and get raw (stripped from HTML tags) text back;
- determine whether the submitted document is a legal document – a neural network is used for this purpose;
- determine whether the document formatting is “broken,” e.g., pasted from a PDF – a neural network is used for this purpose.
Synch LegalNLP is useful when it comes to processing legal text data, especially for ML purposes. For example, to train an AI model (neural network) on a contracts dataset, one needs to “clean” the texts first, which often involves tokenizing different types of information (e.g. replacing all emails or dates with a single token) and stripping the text from HTML tags (e.g. removing “bold”, “underline” and other formatting) to make the AI learning process easier. Synch LegalNLP can take care of such tasks, thus reducing the development time.
Synch LegalNLP was built internally by Synch lawyers-developers.