Decision support system, dedicated to the legal field

A legal decision is a complicated process that requires a thorough analysis of the laws and corresponding decisions in previous similar situations. Legal institutions have vast deposits of information that serve as a basis for knowledge and reasoning. However, according to experts in the field, the analysis of this data is very time-consuming.

A decision support system uses a set of techniques that allow a given user to make a better decision, based on the information available. These techniques come from different fields such as data mining and artificial intelligence.

In a Business Intelligence (BI) context for example, information for strategic decisions is usually not available explicitly. In this situation, data mining is applied to extract informational patterns. More concretely, it consists of applying statistical methods to all available indicators (customer returns, number of purchases, turnover, etc.) in order to identify similar cases in the existing data and thus help make a decision.

In the legal domain, the decision is not information to be predicted, but rather it is made according to well-defined laws. Therefore, we have a set of constraints to make a decision reasoning. The Semantic Web, based on some artificial intelligence technologies, offers a set of tools to build such a system.

The Semantic Web in the legal domain

Since its emergence, the Semantic Web has been adopted by several legal institutions around the world. Indeed, its use has contributed to solving certain problems related to interoperability and documentary research. More concretely, researchers have focused on the development of legal ontologies that use a standardised model to facilitate the exchange of data between different actors. In addition, these ontologies have a set of constraints to answer imprecise queries (e.g. finding all situations similar to this case or based on which laws this decision was made). Recent work in this area includes the LKIF and MetaLex ontologies

The ontology-based information retrieval system can also use descriptive metadata (e.g. user, date, location, etc.) to improve the search result. However, the most difficult task is the development of this ontology and especially the automatic capture of concepts from plain text. Nevertheless, some interesting works exist in the literature and are already being used in the legal domain. As an example, in the context of the development of the Judo ontology, researchers have developed some tools to extract concepts from legal texts through the use of some Automatic Language Processing techniques (e.g. detection of a decision, date, chamber, etc).

Use of constraints for decision support

The use of ontologies in the legal field has certainly played a significant role in the development of the legal field. However, the role of the Semantic Web goes beyond interoperability and document retrieval. An ontology with a set of rules constitutes a knowledge base for developing a decision support system. Indeed, this support system for legal professionals allows them to respond to a given situation, based on two categories of information:

• Strict constraints well framed by law: A legal decision must be based on articles.

• Probabilistic constraints: depending on the contextual elements, the judge must adapt his decision. This information is based on previous decisions.

At ContentSide, we believe that a reliable legal aid system for specific legal situations must take into account the two informational dimensions mentioned above. Namely, constraints and previous decisions. Based on these two dimensions, the system provides a set of recommendations to lawyers to help them make their decisions. We introduce several points of innovation to address some of the challenges encountered, such as

• Scaling up: this is one of the major challenges of a decision-making system. For example, in civil and commercial matters, more than 2 million legal decisions were taken in 2014. In this context, the use of a reasoning system optimised with the exploitation of Big Data technologies becomes essential.

• In the legal field, a decision or law can change over time. This creates a notion of temporality that the system has to deal with.

• The development of a system on an international scale has to cope with the semantic and linguistic heterogeneities between the different laws.

The coupling of data analysis and constraints is certainly a decision-making tool with great potential that can be beneficial to the judicial system. That being said, the use of such a system is not limited to the legal domain. Indeed, many other fields can benefit from such a system (energy, economic, etc.).