uAIS: An Experience of Increasing Performance of NLP Information Extraction Tasks from Legal Documents in an Electronic Document Management System

Nowadays, the huge number of documents which are managed through document management systems make their manual processing practically impossible. That is why the use of natural language processing subsystems that help to perform certain tasks begins to be essential for many commercial systems. Although its use is gradually extending to all levels, this type of subsystems presents the problem of its high requirements of resources from CPU and memory that can harm the entire system to which it intends to provide assistance. In this work, we propose and study an architecture based on microservices and message brokers which improves the performance of these NLP subsystems. We have implemented our approach on a real document management system, which performs intensive processes of language analysis on large legal documents. Experimental results show promising results, greatly increasing the productivity of systems based on other approaches.