Daniel Roßner

SPORE: A Storybreaking Machine

Keywords:

Abstract

We want to present SPORE, a Spatial Recommender System. As we enter a period of unprecedented collaboration between authors and computers, where artificial intelligence in particular seems likely to act increasingly in a co-authoring capacity, SPORE offers a different approach to collaboration. More organic and exploratory than other automated or procedural systems, SPORE aims to mimic the process of storybreaking that already exists in the creative industries.

Demo

SPORE allows its users to add tropes, names, or other ideas to a 2D space. If the system is already aware of the concept, it gives and composes recommendations in the space. In the following image, the user added “Cinderella”.

The application is collaborative, and can be used by multiple users simultaneously. On the left hand side resides the info bar, where users cann add tags and comments to concept.

Recommendations can be replaced with others, or can be accepted by hitting the ‘+’-Button. They become part of the user-controlled concepts and can be re-arranged by the user.

Which recommendations are shown and how they are composed is to a large extend controlled by a visual parser. A software module that interprets human-generated visual structure. If the parser detects groups of objects which belong together visually (by color, position, etc) it issues a query to a knowledge base and receives recommendations.

While takling about ‘knowledge bases’. For this demo we used a versatile algorithm to mine Wikipedia sources. For details, give the paper (6 pages) a read! :) The follwing image gives an overview..

.. or checkout a sample excerpt for ‘stepmother’:

What’s next? Users (authors) may (collaboratively) craft story pieces within the space. The system provides them recommendations to foster the creative process, probably triggering serendipity. Furthermore, the results of the ongoing visual parsing are used to improve the knowledge base. Fun story:

We demoed the system to children. One of them added “seven-headed dragon” to the workspace, thus as part of the story. The system “learned” that concept and connected it to other concepts that where visually related. This way, other children which were testing the system afterwards added this dragon to their story as well, because it was a recommendation.

What would a demo be, without any AI? Of course we did. On the lower right, the application offers a button to “create a fairy tale”. This sends the content of the workspace, along with its parsed visual structure and a specific instruction to ChatGPT. It is primed to create fairy tales with a happy ending and a positive message. The next video shows a small sample (read the result here)

Other Use Cases

e.g. plan your trip to the best conference:

or browse a knowledge base about scientific (hypertext) literature: