Get information from linked data

This demo is also available as a slackbot. See the slack bot source code.

Ask questions to DBpedia. Currently you can ask about the birthday of people and the population of places - more to follow! You will get a response below. The speech input only works with the Chrome browser.

Background

This demonstration works as follows:

• The input (as text or speech) is send to an NLU tool. Here we are using API AI.
• The tool has been prepared to differentiate between uster intents. Currently it just knows two intents: a user wants to look for birthdays of persons or for population of cities.
• The tool differentiates the intents and then detects slots in the user input, e.g. the name of a person or the name of a city.
• The intent filled with slots is transformed to a linked data query. As of writing, the tool uses DBpedia.
• The output of the linked data query is given as textual output or (if supported by the browser) via audio.

The purpose of this tool is to show how workflows with natural language understanding tooling can ease the access to (open or internal) linked data sources. Feedback is very welcome, please send to Felix Sasaki.

Update: now also allows mockup queries to company internal data bases - aks a question about qualifications and you get counts of knowledgeable people.