acter

Acter is a deep learning based solution with legal research and prediction to create legal strategies for lawyers and law firms. Legaltech market is growing and consume a huge amount of investments. 713% growing or 3B$ spending are impressive numbers.

The term, legal research means working with different kind of data sources like texts and cases to mine legally relevant pieces of information, just like Ross or IBM Watson do. Even though everybody talks about Natural Language Processing when it comes to legaltech problems, that works properly only on general language problems. We are seeing the question from a different angle since we are deep learning developers who practice law at the same time.

Natural Language Processing, NLP is a general term in artificial intelligence practice. There are two major types of NLP tasks. The earliest methods used if-then decisions to categorize the text. Better algorithms use statistical models. It means working with words and processing texts based on contexts and usage frequency of words without understanding its meaning. Nowadays NLP is a widely researched topic and there are a lot of AI services out there which use NLP modules. Even though NLP scientists and developers deserve a high appreciation from humanity, this field of deep learning has its own limits. Resolving a legal problem based only on NLP can have the same result than resolving the same problem with a human who can just read the words in the relating texts. Therefore we came to the conclusion, the problem of legal case research and prediction needs a new approach.

The core of our system is a network of legal definitions. Legal definitions are similar to the entries in legal dictionaries. The classical dictionary is static, it contains a lot of titles with definitions. A lingually unified dictionary is good if anybody wants to work with other legal systems or wants to understand the perspective of another contractor. However, to solve a legal problem it is essential to know what are the proper legal definitions and how they work in practice.

Definitions on their own would not be enough to help the computer to properly understand the systematics of the legal cases and legal definitions, therefore we introduced legal operators. These operators connect different definitions together and are able to build up the structure of whole legal cases and definitions in a way computers could understand. Aside from law legal operators come from the field of mathematics and IT to describe the whole network. With the help of the operators a case is understandable for artificial intelligence. Computers couldn’t understand words, they can work with numbers and operators only.