Italian News Agency Selects Expert System for Semantic Software

By: E&P Staff

Agenzia Giornalistica Italia (AGI) has selected Cogito from semantic software provider Expert System, headquartered in Modena, Italy, to more efficiently manage the 50-year-old national press agency’s news content. Based in Rome, AGI has 16 bureaus around Italy and another in Brussels.

Capable of searching discovering, classifying and interpreting text, the Cogito semantic platform enables AGI to reduce its newsroom workload by automatically classifying more than 1,000 daily news items in real time while meeting specific IPTC criteria, enabling fast distribution worldwide. With the technology, AGI can optimize both the general organization of content and the later recovery of the information.

Calling the adoption of the semantic platform “a new and important frontier in technology,” AGI President and CEO Daniela Viglione said in a statement: “Among the new tools now available is the interactive graphic in the Topics of the Day section of our portal,” where readers can “immediately establish logical connections” between events within a net of information.

By understanding the significance of words and phrases, Cogito allows AGI to: identify topics in each article and precisely assign one or more categories to every type of text, selecting the most relevant among the IPTC 1,200 categories; automatically identify the most relevant concepts and specific information used for analysis (semantic tagging) inside each news item; automatically associate news items on similar topics with correlating algorithms, to be able to use semantically similar information (categories, entities, opinions, etc.); and visualize the content of the most recent news using an interactive graphic that links names of people, places, organizations and relevant discussions in the same news item.

Given the publishing industry’s “massive quantities of content,” said Expert System USA CEO Brooke Aker, “Expert System alleviates the bottlenecks created by several different systems working together with competing categorization methods – automatically eliminating this as a separate step and increasing the flow of information.”

Follow by Email
Visit Us

Leave a Reply

Your email address will not be published. Required fields are marked *