Chapter 1: Introduction Modeling Players Generating Content (Chapter 5) (Chapter 4) Game AI Panorama (Chapter 6) Frontier Game AI Research (Chapter 7) AI Methods (Chapter 2) Playing Games (Chapter 3) Chapter 2: AI Methods Power pill eaten Seek for pellets No visible ghost Ghost on sight Chase Ghosts Ghosts flashing Evade Ghosts Seek for Pellets Until ghost on sight Move (Priority) Ghost free corridor Pellet Found Corridor with pellets Eat next pellet Corridor without pellets Attack Enemy Until Health = 0 Spot Enemy 0.5 Mini Gun Select Weapon (Probability) 0.2 Pistol Aim 0.3 Rocket Launcher Shoot! Min Max 5 5 0 10 10 5 -1 5 3 0 2 0 3 -1 0 0 0 1 1 1 0 1 0 0 1 0 0 0 1 0 1 1 0 1 0 1 0 1 0 0 0 1 1 1 0 1 0 0 1 0 0 0 1 0 1 1 0 0 1 1 0 1 0 1 0 p 0 0 1 1 1 0 1 0 0 1 0 0 1 0 1 0 1 1 0 0 0 1 0 0 1 1 1 0 1 0 0 1 0 0 1 0 1 0 1 1 0 0 0 1 π₯2 5 1 4 3 2 6 π₯1 Age? <20 Employed? no No Car >28 [20-28] Salary? Economy yes City Car low Compact fair Sports high SUV Ghost not visible near Pellet Power pill near Aim for pill far Evade ghosts near Aim for pellet fair Aim for pellet far Aim for fruit π₯2 w π₯1 π₯1 π₯2 π₯π Neuron π€1 π€2 π xβw +π π€π π π(x β w + π) π€14 π₯1 π₯2 π₯3 4 π€49 1 π€48 5 8 π8 6 9 π9 2 3 π€37 Input Layer π€79 7 Hidden Layer Output Layer Selection Expansion Tree Policy Simulation Default Policy Backpropagation G Agent State (s) Reward (r) Environment (e.g. Maze) Action (Ξ±) 1 π€01 π€02 π€13 π€14 π€15 π€0π 2 π€01 π€02 π€13 π€14 π€15 π€0π P π€01 π€02 π€13 π€14 π€15 π€0π Fitness value π2 Reward Convolution Action Pooling Chapter 3: Playing Games Win Experience Player Non-Player Motivation Games as AI testbeds, AI that challenges players, Simulation-based testing Motivation Playing roles that humans would not (want to) play, Game balancing Examples Examples Board Game AI (TD-Gammon, Chinook, Deep Blue, AlphaGo, Libratus), Jeopardy! (Watson), StarCraft Rubber banding Motivation Simulation-based testing, Game demonstrations Motivation Believable and human-like agents Examples Examples Game Turing Tests (2kBot Prize/Mario),Persona Modelling AI that: acts as an adversary, provides assistance, is emotively expressive, tells a story, β¦ Imperfect Information Perfect Information Observability Pac-Man Atari 2600 Checkers Chess Go Ms Pac-Man Ludo Monopoly Backgammon Time Granularity Super Mario Bros Halo StarCraft Battleship Scrabble Poker Real-Time Turn-Based Deterministic Non-deterministic Stochasticity Chapter 4: Generating Content 12% 37% 40% 2% 6% 3% Art Manufacturing Other Debugging Marketing Programming Content Method Content Type Necessary Content Stochastic Determinism Deterministic Non-Controllable Controllability Controllable Constructive Iterativity Generate-and-test Autonomous Autonomy Mixed-Initiative Experience-Agnostic Experience Experience-Driven Role Optional Content Generation 0 Midpoint displacement Generation 1 Generation 2 Generation 3 Final Generation Initialize Corner Values Perform Diamond Step Perform Square Step Perform Diamond Step Perform Square Step Experience Agnostic Experience Driven Player (Experience) Super Mario Bros (Pedersen et al., 2010) Sonancia (Lopes et al., 2015) SpeedTree (IDV, 2002) StarCraft Maps (Togelius et al., 2013) Autonomous Sentient Sketchbook (Liapis et al., 2013) Garden of Eden Creation Kit (Bethesda, 2009) Ropossum (Shaker et al., 2013) Tanagra (Smith et al., 2010) Mixed-Initiative Designer (Initiative) Chapter 5: Modeling Players Input Computational (Player) Model Output Gameplay Model-Based [Top-Down] Objective (Psychology, Cognitive Science, Game Studies, β¦) Free Response vs. Forced Response Context Player Profile First Person vs. Third Person Discrete vs. Continuous Numerical (Interval) Regression Time-Discrete vs. Time-Continuous Nominal (Classes) Classification Pre vs. During vs. Post Ordinal (Ranks) Preference Learning No Output Unsupervised Learning (Clustering, Frequent Pattern Mining) Model Free [Bottom-Up] (Data Science, Machine Learning) Activation ( + ) Arousal Fear Excitement Anger Frustration Happiness Valence Pleasant ( + ) Unpleasant ( - ) Sadness Relaxation Boredom Deactivation ( - ) Tiredness Challenge Flow Channel Anxiety Boredom Skills Nearest Monster Nearest Treasure Monster Nearest Treasure Safe Treasure Nearest Portion Safe Nearest Portion Treasure Safe Portion Exit Portion Safe Hit Points Exit Safe Exit Exit Safe Chapter 6: Game AI Panorama DO (Process) WHAT (Context) FOR WHO (End User) Designer Model Content Player AI Researcher Generate Behavior Producer / Publisher GAME AI AREA Generate Content (Assisted) Model Players (Behavior, Experience) Generate Content (Autonomously) Play Games (Win [NPC], Experience [NPC]) Play Games (Win [PC], Experience [PC]) Model Players (Behavior) Player Interaction Model Players Experience Behavior Game Generate Content Autonomously Content NPCs Play Games [as NPC] Win Experience Play Games (as PC or NPC) To Win Model Players Experience Behavior For the (Game) Experience Generate Content Assisted Autonomously Play Games (as PC or NPC) To Win Model Players Experience Behavior For the (Game) Experience Generate Content Assisted Autonomously Play Games (as PC or NPC) To Win Model Players Experience Behavior For the (Game) Experience Generate Content Assisted Autonomously Play Games (as PC or NPC) To Win Model Players Experience Behavior For the (Game) Experience Generate Content Assisted Autonomously Play Games (as PC or NPC) To Win Model Players Experience Behavior For the (Game) Experience Generate Content Assisted Autonomously
© Copyright 2026 Paperzz