Artificial Emotional Intelligence in Virtual Creatures

Introduction

Nowadays the connection between a Human being and computer based systems is growing stronger. Many innovations appear, for instance, robot toys like Pleo, Robosapiens or Honda Robot, and even conventional computers with the ability to control complex systems, such as virtual reality worlds. This phenomenon shows that human-computer interaction (HCI) is deeper than just the delegation of work to the machine. This remark makes evident two fields of research. The first one is the physiological field, and the second one is the psychological field that proposes models that represent emotions in computer based systems. In this work we deal with the second field of research.

Today a large number of studies about emotions exist and there are several models of emotion proposed in the literature, unfortunately we cannot know for sure which models are correct. In this chapter we deal with the problem of expressing corporal emotions. The main issue is that expressing emotions is completely subjective for everyone. This fact makes very difficult the establishment of general models for emotions. An alternate approach explored in this chapter consists in designing a model based on some observations on the behavior of Human being. Psychologists traditionally single out three types of human intelligence: Abstract Intelligence that denotes the ability of Human being to understand and manipulate verbal and mathematical symbols, Concrete Intelligence which indicates the capacity of Human being to understand and manipulate objects, and Social Intelligence that allows the Human being to understand other individuals and to interact with them. The EI of Human being has its roots in the Social Intelligence (SI), which is divided in two categories: Interpersonal Intelligence as ability to understand other people and Intrapersonal Intelligence as self-consciousness.

Emotions are an important aspect in the functioning of the human mind. Nevertheless, the role that emotions play in our actions, behavior and thinking has been misapprehended and misinterpreted. The old philosophers did not consider the emotions as an important aspect of human intelligence, by the contrary they perceived the emotions as an impediment that blocks and prevents the human reasoning and thought. In the Plato’s Phaedo dialogue, Plato explained that fears, passions and desires make thinking and reasoning impossible. Later, Descartes based on the same idea his defining of emotions as passions or needs that the body imposes on the mind.

Recently, several psychologists have begun the exploration and study of emotions to explain better their functioning, which is an important component of the human intelligence and cognition. The obtained results give evidence that emotions have an important impact on thinking, judgment, reasoning, memory and decision making of Human being. Gardner (1983) introduced the term of Multiple Intelligences for describing the personal intelligence as a type of human intelligence that includes social interactions and emotions. Damasio (1994; 1995) demonstrated in neurological studies that people who lack the capacity of emotional response can take incorrect decisions and execute mistaken actions that can limit their performance in society. Basing on the fact that emotions are an important part of human intelligence, Goleman (1995) coined the concept of EI.

Many psychological models have been proposed to describe the emotional functioning of the human brain and the mental processes. Several models are centered on the effect of motivational states or on the processes by which the events trigger certain emotions. Such models are called event appraisal models. Other models examine the influence of expectations on emotions. But none of these models presents complete abstraction and shows general idea. In these models the emotions are considered to be mental states generated with the use of mapping that includes great variety of considerations, such as expectations, motivational states, events, and environmental conditions.

Inspired by the psychological models of emotion, many researchers have recognized in Artificial Intelligence (AI) the importance and utility for improving complex, dynamic and interactive virtual environments with the help of the computational models of emotions. Designed models of emotions can represent a better understanding of moods and emotions and can thus adapt to needs of program or application. However, most of designed computational models of emotion only represent specific situations and respond in predetermined way to them.