Emotionally Driven Natural Language Generation for Personality Rich Characters in Interactive Games

Abstract

Natural Language Generation for personality rich characters represents one of the important directions for believable agents research. The typical approach to interactive NLG is to hand-author textual responses to different situations. In this paper we address NLG for interactive games. Specifically, we present a novel template-based system that provides two distinct advantages over existing systems. First, our system not only works for dialogue, but enables a character's personality and emotional state to influence the feel of the utterance. Second, our templates are resuable across characters, thus decreasing the burden on the game author. We briefly describe our system and present results of a preliminary evaluation study.

Plan-Based Character Diversity

“…Complex character modeling has been explored to such a great extent that, rather than attempt exhaustiveness, we will cite several illustrative examples. NPCs have been modeled in terms of personality traits (McCoy et al, 2010), emotional states (Cavazza and Charles, 2005;Strong et al, 2007;Cavazza et al, 2009), relationships (Cavazza and Charles, 2005;McCoy et al, 2010;Thue et al, 2010), andneeds (McCoy et al, 2010;Paul et al, 2010), among others. In the related field of behavioral robotics, emotions are modeled for Sony's AIBO robots Arkin et al (2003).…”

“…In the related field of behavioral robotics, emotions are modeled for Sony's AIBO robots Arkin et al (2003). Substantial work deals with the complexities of generating interesting and realistic language/dialogue that is both character-and contextdependent (Cavazza and Charles, 2005;Strong et al, 2007;Lin and Walker, 2011). Another significant area of focus is NPC behavior in the context of storyline consistency and variety (Cavazza, Charles, and Mead, 2002;Porteous, Cavazza, Charles, 2010;McCoy et al, 2010;Thue et al, 2010).…”

“…Diversity can be obtained by modeling characters in terms of personalities, morals, emotional states, and social affinities. Such character modeling has been explored extensively (Cavazza, Charles, and Mead, 2002;Cavazza and Charles, 2005;Strong et al, 2007;McCoy et al, 2010;Cavazza et al, 2009;Thue et al, 2010).…”