Webinar: Early Scientific Intelligence Pipeline - Sten Christensen & Brian Michels
Sten Christensen and Brian Michels described a project to increase the pipeline of early scientific intelligence about opportunities for innovation from external collaboration and partnerships; and to enable scientists to exploit these early signals. The informatics challenges were covering more data sources, filtering the information, and matching it with what was already known; a technical challenge was to combine the Linguamatics NLP platform with Novo’s InfoDesk system.
Novo’s information sources were extended to include Medline, patents, NIH grants, Dow Jones data, insight memes, conference data, and technology transfer news. These sources are queried using the Linguamatics NLP platform to extract key relevant data, and the Linguamatics query output is fed to the InfoDesk for consolidation. Sten and Brian described the technology behind the Linguamatics/InfoDesk connection in some detail. These results are then manually validated and curated by subject matter expert curators before they are disseminated to the scientists.
In closing, Sten and Brian identified three areas for more attention: automating processes; encouraging scientists to be more open on a chat system like Yammer; and improving techniques for finding information about start-up companies in sources like Twitter.