Essays.club - Get Free Essays and Term Papers
Search

User-Based Character Modelling for Interactive Storytelling

Autor:   •  March 10, 2018  •  7,702 Words (31 Pages)  •  494 Views

Page 1 of 31

...

accuracy is not one of the objectives of an entertainment application, an even better approach is to use some short version of such tests, where we lose in precision but gain in a smaller demand of time from the user [Gosling et al. 2003].

In this paper we present the development of a prototype that permits the modelling, by the user, of some characters’ personality traits in an interactive storytelling application. The paper is organised as follows. Section 2 presents related works dealing with personality traits and the application of user modelling in interactive storytelling systems. Section 3 shows the decision-making process this work is based on. Section 4 describes the steps of the character modelling process and shows how it can help improving the diversity of generated stories. Section 5 presents a small evaluation through some examples. Finally, in section 6 we discuss the concluding remarks.

2. Related Work

Some recent research has addressed the usage of user models in interactive storytelling, particularly with the help of stereotypes [Rich 1979] to represent users’ preferences. El-Nasr [2004a][2004b] presents an interactive story called Mirage, that makes use of player knowledge to enforce user engagement. This task is accomplished by borrowing the definition of its characters behaviours from acting theory, along with inferred knowledge about player actions. Based on predicted player behaviour, the actors choose between different tactics [El-Nasr 2007] to try to reach their goals for a certain scene; when a goal is not reached by a certain behaviour, another approach will be selected by the actors so that they can continue trying. The dramatisation of Mirage’s interactive narrative happens in a rich 3D world supported by an architecture that implements an agent model and utilises varying animations attached to the same action with different “adverbs” associated to them, i.e. if a character wants do draw a sword, it can do it slowly, violently, etc. [Karlsson et al. 2009]. Mirage also makes use of a scripting language that allows designers to define an evaluation function that influences the way the system estimates a user’s character given its actions and story context. To represent a character’s personality, Mirage makes use of a vector of stereotypes based on the following five dimensions: heroism, violence, self-interestedness, truth seeking, and cowardice.

Another interactive storytelling system that uses stereotypes is PaSSAGE (Player-Specific Stories via Automatically Generated Events), which uses automatic player modelling to learn the player’s preferred style of play [Thue et al. 2007]. This model is then dynamically used to select the content of an interactive story from a library of possible events, called encounters, each having been annotated by an author with information concerning which player types it would be suitable for [Karlsson et al. 2009]. When determining which encounter to run, PaSSAGE examines each encounter’s set of branches. To help maintain a stronger sense of story, encounters are grouped into sets corresponding to the many phases of the Monomyth [Campbell 1949].

These encounters can be refined and are implemented by the use of triggers, usually started when a player approaches some suitable location. Characters satisfying the encounter’s trigger conditions assume the behaviours authored for this event, which are tailored to encourage the player’s preferred styles of play. The player model vector then changes depending on player action selection. The main difference from the stereotypes in PaSSAGE to the ones in Mirage is that Mirage tries to model the player’s character, defining its model as values along traits of character stereotypes. On the other hand, PaSSAGE categorises player type stereotypes based on a set of player types as published by Robin Laws [2001]. These player types include Fighters (who prefer combat), Power Gamers (who prefer gaining special items and riches), Tacticians (who prefer thinking creatively), Storytellers (who prefer complex plots) and Method Actors (who prefer to take dramatic actions).

More recently, Lima & Silva [2010] also proposed the usage of viewers’ stereotypes in a multimodal version of the interactive storytelling system Logtell [Ciarlini et al. 2005]. Logtell utilises goal-inference rules, automated planning and user intervention to dynamically create plots that are dramatised with 3D graphics (Figure 1). New events are suggested to the viewers to be incorporated to the story, and they can accept or reject them either by voice (saying “yes” or “no”) or by body gestures (nodding or shaking the head for acceptance or rejection, respectively). Viewers can also choose not to interact; in this case, the suggestion is simply discarded. The viewers’ choices are recorded and the system can evaluate how much they interact, classifying them into one of three viewers’ stereotypes: active viewers (that like to interact with the story), conventional viewers (that like to interact only in key points) and passive viewers (that only like to watch the story without interaction). The system can also adapt itself to increase or decrease the level of suggestions presented to the viewers according to their stereotypes. These stereotypes can change during dramatisation sessions as long as the viewers’ behaviour change.

Figure 1. A scene during the dramatisation of a plot in Logtell

Besides the quantitative measure concerning the level of intervention made by the viewer, another information that can be extracted from the viewer’s history is the qualitative aspect of the choices they made, i.e. the characteristics of the events that were chosen to be incorporated to the story. The idea is to enhance the viewers’ experience by capturing their narrative preferences (regarding the events to be incorporated to the story). To accomplish this, the events used in the storytelling context adopted in this version of Logtell (a small sub-class of the popular Swords & Dragons genre) have been described regarding its atmospheric traits. There are five such traits (adventurous, romantic, magical, violent, and dark) indicating the “feelings” that those events can be considered bringing to the story.

The recorded history of the choices made by the viewers for every interaction presented is used to find the average atmospheric traits preferred for each viewer. Likewise, a generic model of preferences regarding these traits is created for the whole group of viewers, considering all the interactions of all the users. These values are then compared to the viewers individually in order to find possible viewers that deviate very much from the rest of the group. This information

...

Download:   txt (50.2 Kb)   pdf (107.6 Kb)   docx (37.2 Kb)  
Continue for 30 more pages »
Only available on Essays.club