|
Menu |
|
|
Projects |
|
|
News - Highlights |
|
|
Member Area |
|
|
|
|
Computational Narrative Blending
|
|
|
|
Research Project
Informatics Department - Pontifical Catholic University of Rio de Janeiro (PUC-Rio)
IADE - Universidade Europeia
Edirlei Soares de Lima <edirlei.lima@universidadeeuropeia.pt>
Bruno Feijo <bfeijo@inf.puc-rio.br>
Antonio L. Furtado <furtado@inf.puc-rio.br>
Narrative blending the process in which two or more narratives are blended together to generate a new narrative variant. This concept comes from notion of conceptual blending (Fauconnier and Turner, 1998; Fauconnier and Turner, 2003), which has been proposed as a fundamental cognitive process where two or more conceptual spaces are merged to form a new blended space. In this project we explore conceptual blending in a narrative domain and propose a new plot generation method based on the reuse of existing stories. Starting from a set of stories sharing the same narrative structure, our method is capable of generating new narrative variants that combine episodes extracted and adapted from different stories. By combining a plan validation algorithm with a basic narrative structure, our method guarantees the logical coherence and general structure of the generated narratives.
Publications:
|
Lima, E.S., Feijo, B., Furtado, A.L., 2021. Computational Narrative Blending Based on Planning. Proceedings of the 20th IFIP International Conference on Entertainment Computing (ICEC 2021), Coimbra, Portugal.
|
Prototype Application:
In order to apply and evaluate the proposed method, we implemented in C# a tool to assist amateur/professional authors in the process of visualizing all narrative variants that can be create for a set of existing stories, which be used as a source of inspiration for new story ideas. Besides applying the proposed narrative blending method, the tool also permits users to interfere in the initial state of the narratives by changing existing facts (i.e. changing the names of characters, places and objects) and by adding new facts (i.e. creating new characters, places, objects and relations). Such interference extends the capacities of the proposed method by allowing it to generate even more diversified narratives that would not be logically possible just with the original states of the narratives stored in the domain library. In addition, it allows the generation of more personalized narratives (for example, allowing the author to decide who will play the role of hero in the stories). Users can define the customized initial state by manually writing logical facts or by using a state design tool that allow them to create characters, objects, and stablishing relations between them visually.
Download:
|
|
|
|
|
|
|
