Jump to Navigation

Modelling a conversational storyworld

Conversational Storyworld BasicsScenejo System Description

Conversational Storyworld Basics

This approach is demonstrated on the Scenejo system and its authoring tool SAT.

Scenejo System Description

Scenejo, as a chat bot based system, makes use of so-called stimulus response elements (SRE). These smallest structural elements of the system are grouping the input and output of a virtual character. The input is what it “hears”, its stimulus, and the output what it will “answer” on that certain input, the response. The system also offers an abstract layer above the final wording level to group utterances of the same meaning or – the other way round – to offer alternative utterances in the same context. There can not only be multiple different final outputs for an abstract output, but also multiple abstract outputs for an abstract input, which are chosen depending on preconditions.

The two pictures below are showing two different kinds of visualisations of the conversational stories. The first one shows the 3d realtime rendering of the characters as floating heads. The sound is generated by speech synthesis. The second one shows the video client version which plays videos according to the utterance. The audio is prerecorded and added to the video material.

Scenejo01 Scenejo02

The picture below shows the "Scenejo Authoring Tool" (SAT).

Scenejo03




by Dr. Radut.