New version page

GT LCC 3710 - Artificial Life Meets Entertainment

Upgrade to remove ads
Upgrade to remove ads
Unformatted text preview:

The relatively new field of artificial life attempts to studyand understand biological life by synthesizing artificiallife forms. To paraphrase Chris Langton, the founderof the field, the goal of artificial life is to “model life asit could be so as to understand life as we know it.” Arti-ficial life is a very broad discipline which spans suchdiverse topics as artificial evolution, artificial ecosys-tems, artificial morphogenesis, molecular evolution,and many more. Langton offers a nice overview of the different researchquestions studied by the discipline [6]. Artificial life shares with artificialintelligence (AI) its interest in synthesizing adaptive autonomous agents.Autonomous agents are computational systems that inhabit some complex,dynamic environment, sense and act autonomously in this environment, andby doing so realize a set of goals or tasks for which they are designed.The goal of building an autonomous agent is as old as the field of AI itself.The artificial life community has initiated a radically different approach tothis goal, which focuses on fast, reactive behavior, rather than on knowledgeand reasoning, as well as on adaptation and learning. Its approach is largelyinspired by biology, and more specifically the field of ethology, whichattempts to understand the mechanisms animals use to demonstrate adaptiveand successful behavior. Autonomous agents can take many different forms, depending on thenature of the environment they inhabit. If the environment is the real phys-ical environment, then the agent takes the form of an autonomous robot.Alternatively, one can build 2D or 3D animated agents that inhabit simulat-ed physical environments. Finally, so-called knowbots, software agents orinterface agents are disembodied entities that inhabit the digital world ofcomputers and computer networks [8]. There are obvious applications forall these types of agents. For example, autonomous robots have been builtfor surveillance, exploration, and other tasks in environments that are unac-cessible or dangerous for human beings. There is a long tradition of build-108 November 1995/Vol. 38, No. 11 COMMUNICATIONS OF THE ACMArtificial Life MeetsEntertainment: LifelikeAutonomous AgentsPattie Maes The potential of AIin the field ofentertainment isslowly but steadilybeing unearthed ina series of new, cre-ative applicationsthat reflect theeven grander futureof artificial life.AIemerging technologiesing simulated agents for training purposes. Finally,more recently, software agents have been proposed asone mechanism to help computer users deal withwork and information overload [8].One potential application area of agent researchthat has received surprisingly little interest so far isentertainment. This area may become much moreimportant in the coming years, since the traditionalmain funding source of agent research, the defenseindustry, has been scaling down. Entertainment is anextremely large industry that is only expected to growin the near future. Many forms of entertainment fea-ture characters that act in some environment. This isthe case for videogames, simulation rides, movies,animation, animatronics, theater, puppetry, certaintoys, and even party lines. Each of these entertain-ment forms could potentially benefit from the castingof autonomous semi-intelligent agents as entertain-ing characters. Entertainment is a fun and very chal-lenging application area that will push the limits ofagent research. The Challenge of ModelingEntertaining CharactersSeveral forms of commercial entertainment currentlyincorporate automated entertaining characters. Mostof these characters are extremely simple; theydemonstrate very predictable behavior and do notseem very convincing. This is particularly the case forcharacters with whom a person can interact in real-time, for example, videogame characters. When auto-mated characters show sophisticated behavior, it istypically completely mechanical and noninteractiveand the result of a painstaking and laborious process.An example of the latter is the behavior of thedinosaurs in the movie Jurassic Park. COMMUNICATIONS OF THE ACM November 1995/Vol. 38, No. 11 109Space-Ace says, “So why won’t you kiss me, Julia?”Julia says, “Sorry, I’m not into geeks, Space-Ace.”Space Age says, “Julia, I’m not a geek.”Julia says, “Then lose the pocket protector, Space-Ace.”Space-Ace says, “Julia, I don’t have a pocket protector.”Julia says, “Congratulations.”Space-Ace says, “So kiss me, Julia.”Julia says, “I’d rather be overhauled by a Yugo mechanic, Space-Ace.”Space-Ace says, “julia, kiss me”Julia says, “Why, did your sheep leave you?”Space-Ace says, “julia, kiss me”Julia says, “Take a long walk through an unlinked exit, Space-Ace.” (Copyright © 1990 byAshne)Julia is offended by Lenny’s suggestion.Julia winks at Lenny.Julia eyes Lenny carefully...Julia grins at Lenny.Julia nods to Lenny.[...]Figure 1. Realistic fish behavior modeled by Terzopoulos et.al. to produce short animated moviesFigure 2. Tosa’s artificial “baby” reacts to sounds made by auser looking into its cribFigure 3. Excerpts from interactions with “Julia,” Mauldin’sautonomous conversing agent that inhabits atext-based multi-user environmentIn the last couple ofyears, a few exceptionshave emerged. A num-ber of researchers haveapplied agent technolo-gy to produce animatedmovies. Rather thanscripting the exactmovements of an ani-mated character, thecharacters are modeledas agents that performactions in response totheir perceived environ-ment. Reynolds [11]modeled flocks of birdsand schools of fish byspecifying the behaviorof the individual animals that made up the group. Thesame algorithms were used to generate some of thebehavior of the bats in the movie Batman II. Terzopou-los and colleagues [12] modeled very realistic fishbehavior, including mating, feeding, learning, and pre-dation (Figure 1). His models have been employed tomake entertaining, short, animated movies.In addition to the previous work, some researchershave used agent models to build interactive, real-timeanimation systems. Bates’ Woggles World [1] allows theuser to interact with a world of creatures called Wog-gles. In this pioneering work, a user interacts with theworld and its creatures using the mouse and key-board to directly control the movements and behav-ior of one of the Woggles. The different Woggleshave internal needs and a wide range of emotions,which


View Full Document
Download Artificial Life Meets Entertainment
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Artificial Life Meets Entertainment and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Artificial Life Meets Entertainment 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?