A virtual assistant is designed to respond to user queries and enables users to navigate web pages and/or applications. However, full-text pages, e.g., pages of online encyclopedias or long-form journals, are not easy to navigate, especially via a voice interaction. Text is linear, includes reference points provided by authors, e.g., title, summary, etc., and, in raw form, is unsuitable for navigation via voice. Per the techniques of this disclosure, a text page or corpus is transformed, e.g., by the use of relation-extraction technique, into a distilled knowledge representation, e.g., a knowledge graph that includes relates the concepts on the text page with other concepts. A virtual assistant surfaces the distilled contents to the user in a manner that is more natural than a linear readout of the page. Navigation to relevant information is speeded up, and opportunities are offered to the user to explore related content.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.