Turn your knowledge graph into your personal assistant!

metaphactory was extended with an Amazon Alexa voice interface. The generic interface enables end users to interact with any knowledge graph in the metaphactory through natural language.

To demonstrate this new feature, we published an Alexa Skill able to answer factual questions about any kind of entity or concept and their properties described in the Wikidata knowledge graph. Using this Alexa Skill, users can ask their Amazon Echo or Amazon Echo Dot questions such as "What is the birthday of Barack Obama?" or "What is the capital of Lithuania?" through natural language. metaphactory retrieves the answers to these queries from the Wikidata knowledge graph.

Curious? Enable this Alexa Skill and try it on your own Alexa device!

To learn more about this use case, please download our demo paper which was presented at the International Semantic Web Conference 2017.

Relying on SPARQL queries to access the semantic data, metaphactory supports both a single-repository scenario as well a virtual integration scenario: the data from multiple data sources are accessed using a federated SPARQL query processing engine. The platform provides a set of services on top of these integrated semantic data. These services are in turn utilized by the frontend layer. The user interface is based on a customizable templating mechanism which provides various UI elements for data visualization, authoring, and search, configurable as HTML5 components.

The Amazon Alexa framework allows the application engineer to define a specific service (called Skill) which will process user requests targeted at it. To enable accessing semantic data using the skill framework, we developed a mechanism to generate the definition of a skill automatically from a given knowledge graph containing factual information. Through this new service, metaphactory delivers an alternative interaction mechanism alongside more traditional paradigms: voice interaction using natural language questions.