| Resumo: |
Our project proposal lies on the use of algorithms to automate the creation of science news stories from the structured data that can be found in open
repositories of universities. We will be using the University of Porto Open Repository as a data source for the project development. All the
publications deposited in this repository are available in open access and covered by the Creative Commons public license.
The "Algorithmic Science News Agency" will collect available data and process it, identifying data markers as newsworthy elements and arranging
them so that by following a set of rules and templates it will be able to generate short stories, allowing for an increase of the amount of science news available.
The data in these repositories is highly complex and although natural language generation technology is evolving at a fast pace, domain experts will be called to ascertain the newsworthiness of the stories as well as help identify patterns and build different templates for a diverse catalog of science subjects and topics. |