Artificial Intelligence in Computer Games

1
Artificial Intelligence in Computer Games

• Knut Vidar Siem
• City University of New York, College of Staten
  Island
• Department of Computer Science
• for
• Seminar in Computer Science (CSC490-6220)
• Spring 2006

2
Covered Topics

• AI techniques used in games
• Some game genres and how they use AI
• Quick case study Quake III Arena
• AI research and game developing
• Effects of AI in games
• (AI for education and tutoring)

3
AI in a Computer Game Setting

• The look and feel of the AI is far more
  important than its inner workings
• The player is supposed to win
• The overall goal is to create a fun and
  entertaining game
• With this in mind (commercial) game developers
  can include AI in their games

4
AI Techniques in Computer Games

• Finite state machines (FSM)
• Artificial neural networks (ANN)
• Fuzzy logic
• Expert systems
• Genetic algorithms

5
Finite State Machine

• A finite state machine is a model of behavior
  that can be organized in states
• Strengths
• As complex as the desiredcomplexity of the
  subjectbeing modeled
• Weaknesses
• Not very accurate formodeling human behavior

6
Artificial Neural Network

• Models a network of biological neurons
• Each node has input and output
• Strengths
• Generalization
• Human-like solutions
• Weaknesses
• Training takes time

7
Fuzzy Logic

• Derived from fuzzy settheory
• Elements have membershipvalues in the interval
  0, 1
• Introduces the possibilityof imprecision

8
Expert System

• Components
• Knowledge base
• Set of production rules
• Inference engine for reasoning over the knowledge
• Strengths
• Can solve problems requiring expert knowledge
• Weaknesses
• Narrow domain of knowledge
• Inherently just decision logic

9
Genetic Algorithms

• Principles from genetics and natural selection
• Evolutionary techniques are used to optimize the
  solution to a problem
• A population of solutions evolve over generations
  by combining properties (often with the
  possibility of mutation)

10
The Most Popular Game Genres

• There are a wide variety of genres the following
  are the most popular of the major ones
• First-person shooters (FPS)
• Real-time strategy games (RTS)
• Role-playing games (RPG)
• Among others are simulator, sports and fighting
  games
• The AI requirements varies with the genre

11
First-Person Shooters

• Animation control the bodyand make it blend
  with theenvironment e.g. using adecision system
• Movement how the entitymoves in the world with
  pathfinding
• Combat should do spatial reasoning, tactics,
  perceptions and e.g. using FSMs
• Behavior control the entitys state of mind
  using FSMs and fuzzy logic

Quake III Arena, 1999
12
Real-Time Strategy

• Analysis define goals andprioritize through map
  analysisand area labeling
• Resource allocation matchmeans with goals with
  thehelp of network flow algorithms
• High-level AI switch between AI modes using an
  FSM, a rule-based system or a fuzzy logic system.

StarCraft, 1998
13
Role-Playing Games

• Historically not a genrewith a lot of AI
• Random encounters aremore common
• How much AI is useddepends on how rigid
  thestory line is
• The AI techniques used depend greatly on the kind
  of intelligence being modeled.

Final Fantasy XII, 2006
14
Case StudyQuake III Arena

• First-person shooter (FPS) released by Id
  Software in 1999.
• The player moves around in a 3D environment
  fighting enemies and assisting teammates
• In this virtual world lies items such as weapons
  and power-ups
• The players can be either human or computer
  controlled bots

15
Quake III Arena

• "Half way through the project, I think everyone
  said, 'The bots suck. We have to get our act
  together.'"- Christian Antkow, Level Designer
• Their rescue was Jean-Paul van Waveren, a
  22-year-old Master student from the Netherlands

16
Quake III Arena Bots

• are autonomous self-controlled
• use a cognitive model of the world,eliminating
  the need for waypoints
• use an FSM-like structure to think
• have some wired domainknowledge
• can issue and react to teamcommands

Crash
17
Quake III Arena Bots

• use an interface of commands similar to the one
  presented to the player
• depend on fuzzy relations that specifies how much
  they want to do, have or use something (in a
  tree)
• have goals and sub-goals on a stack
• dont do much planning
• do not cheat by accessing more information than
  they should

18
Quake III Arena Bots

• Video from a human viewpoint
• Video from a bots viewpoint

19
Common AI Problems

• Cheating, i.e. peeking into secret information
• Super-powers, more health, resources etc.
• Unless an AI entity is made a bit dumb it appears
  to be a super-human
• Unhandled mistakes look bad (and they always
  happen)

20
Expected AI Qualities

• These qualities are parts of the ultimate goal
  in AI
• Predictability vs. unpredictability
• Creativity in problem solving
• Personality
• Intension and autonomous acting
• Improvising and planning
• Learning

HAL 9000
Data
21
Why Computer Games are suitable for AI research

• Technical reasons
• The virtual environment of a game is not a
  simulation of the problem domain, it is the
  problem domain
• Games free AI researchers from building the
  environment
• They often come with good APIs and modification
  possibilities

22
Why Computer Games are suitable for AI research

• Non-technical reasons
• Games are cheap, often below 50
• Game development is a big, high-paced industry
• As games continue to improve, AI is expected to
  be the next discriminant (as graphics was a few
  years ago)

23
Effects of AI in Games

• A good AI doesnt necessarily make the game fun
• More interesting behavior
• Better adaptability and improvisation
• More human-like flaws (if implemented)
• Further blurring of the border between reality
  and unreality
• What are the ethical issues with AI entities?

24
AI in Educational Tools

• Can reduce cost of expensive training
• Can simplify training operations that are
  difficult to arrange.
• An AI entity can adapt to the progress and attend
  to the specific needs of a student.
• Adds personality, empathy and emotion to the tool
• Programs exist e.g. to teach programmers to
  detect bugs (PROUST) and to prevent bullying in
  schools (FearNot!)

25
References

• Human-Level AI's Killer Application Interactive
  Computer Games, John E. Laird Michael van Lent,
  2001
• Agent design to pass computer games, Astrid
  Glende, 2004
• The Quake III Arena Bot, J.M.P. van Waveren,
  2001http//www.kbs.twi.tudelft.nl/docs/MSc/2001/W
  averen_Jean-Paul_van/thesis.pdf
• Strategies for Strategy Game AI, Ian Lane Davis,
  1999http//www.maddocsoftware.com/pdf/I_Davis_99-
  Strategy.pdf
• Learning by Feeling Evoking Empathy with
  Synthetic, Characters, Ana Paiva, 2005
• Links artificial intelligence and interactive
  entertainment,
• AI in First-Person Shooter Games, John McCloskey,
  Jeffrey Miller, Amish Prasad Lars
  Lindenhttp//ai.eecs.umich.edu/soar/Classes/494/t
  alks/Fps.pdf