The company’s history is in doing custom text analytics solutions for big customers, and the new service brings to a wider base its experience providing application-specific APIs to vertical requirements and industries.
It’s starting with offering as a hosted service packaged, high-level, multi-language semantic capabilities for social and traditional media analysis, as well as its Core API for providing horizontal services such as syntactic category keyword filtering, named entity recognition, and rule-based pattern application. A semantic publishing API is next on the vertical roadmap.
“We are trying to create precision pipelines for their needs, with higher-level functionality targeted for specific markets,” says Antonio Matarranz, who leads marketing and sales for Daedalus. Customers, he says, benefit from its prepackaged and app-specific functionality when it comes to leveraging semantic technology to understand meaning across different scenario types, such as the more formal structure of news articles compared to the more informal speech applied in social media. José Luis Martínez Fernández, who’s in charge of product at the company, says that Daedalus uses a mix of semantic technologies in its offerings, from NLP and linguistic resources to statistical methods, and applies whichever one fits best to the task.
“If you use the same pipeline for social media [as for more traditional media], it doesn’t work because the language is so informal,” says Matarranz. And, he adds, “whether it’s useful to use a general dictionary or taxonomy or to apply a very specific taxonomy or ontology, we have prepackaged all this knowledge and our know-how in our web services.”
Users such as developers of monitoring and analysis tools, PR and communications agencies, and corporate clients can use its API for Media Analysis to identify references, reviews and buy signals in social and traditional media. Telefonica ran a pilot using the SaaS offering to assess its applicability for targeted advertising purposes both in its social network (Tuenti) and smartphone messaging app (TU Me), for which a prototype was built, Matarranz says.
“If you are a developer or integrator and you want to use the semantic functionality, you can use that off shelf,” Matarranz says. “We really are trying to hide the technical complexity of using this technology.”
The dynamic semantic publishing API should debut by September, with functionality for content tagging, enrichment and enabling relations to other sources of content.