Evolving stories: Grammar evolution for automatic plot generation

Abstract

In this paper, we propose a computational framework for automated story-based scenario generation. Under this framework, a regular grammar is developed to model various causal relationships inside a given story world. The grammar is then evolved using evolutionary computation techniques to generate novel story plots, i.e. story-based scenarios. To evaluate these newly generated scenarios, a human-in-the-loop model is used. An experimental study was carried out, in which the proposed framework was used to create novel plots based on the famous Little Red Riding Hood fairy tale. The experimental study demonstrated that evolutionary computation can potentially contribute significantly to story generations. Some challenges were identified including the difficulty to quantify such subjective measures as plot interestingness and creativity.

An expressive GL-2 grammar for representing story-like scenarios

“…We verified the new representation scheme using the scenario evolution framework proposed in [7] to evolve air traffic scenarios with the setting taken from the example scenario of Section IV.…”

“…2) Genetic Operators: Classical crossover and mutation operators can be used as shown in [7]. However, a repair operator is added to the framework to guarantee the validity of each individual chromosome.…”

“…This technique is greatly limited by the number of existing cases in terms of novelty, unpredictability, and uncertainty found in newly generated scenarios. Recently, Bui et al [7] have proposed to use a grammar evolution approach to generate story plots with some degree of coherence, consistency, and interestingness. Nonetheless, this work has some limitations as discussed in the introductory Section.…”

“…The approach can be supplemented with search techniques, e.g. grammar evolution as in [7], [34] to generate useful scenarios.…”

“…Although a notable progress has been made [3], [5], [7], [10], [12], [17], [23], [35], generating stories and story-like scenarios remains a challenging problem. In an earlier work [7], we have made a successful attempt in using evolutionary computation to compose story plots with some degree of coherence, consistency, and interestingness. However, the work had a number of limitations particularly with regard to story representation and story evaluation.…”