SAS and Notilyze collaborated on a pilot project for a Belgian governmental department. A team of legal experts within this department is responsible for responding to policy questions from the Parliament and companies. To guarantee the historical consistency of responses, the legal team members sifted through a database of 30,000 historical questions and answers. Notilyze, SAS and the customer collaborated on the “Smart Search” pilot project to improve the existing Q&A database and search experience. During the pilot project, machine learning and AI techniques were used to automatically extract relevant legal metadata information from documents and to automatically classify legal documents into a granular taxonomy of policy-related topics. In addition, we introduced modern search techniques (TF-IDF, semantic and fuzzy search, and multilingual search) to boost the relevancy of search results. The presenters illustrated how SAS Viya’s machine learning and visualization capabilities, Microsoft Azure’s cloud computing and AI services, and SingleStore’s powerful machine learning database can be combined to provide a next-generation search and information retrieval experience.
Presentation slides are attached to this post.
