Frankencritters

1
Frankencritters

• Greg Reshko and Chris Smoak

2
Background

• 1989
• Larry Yaeger Apple Computer
• Polyworld Artificial Life Software
• Simulated small creatures that could eat, mate,
  attack, see, and move
• 5 - 15 sec./frame
• Some emergent behavior showed promise

3
Artificial Life

• Model and simulate complex biological systems
• Usually combines multiple traditional AI parts
• Introduces more biologically-based parts
• Explore complex systems
• Life, Tierra, Eden, Polyworld, etc.

4
Goals

• Continue Polyworlds intentions
• Improve performance
• Improve algorithms and correctness
• Observe emergent behavior
• Learn about ALife and complex systems
• Validate biologically-based complex systems

5
Simulated World

• Large open space for critters to live in
• Not too large to encourage interaction
• Critters
• 50 100 at once
• Obstacles
• Plants
• Long simulation time

6
(No Transcript)
7
Critter Design - Physical

• Simple triangular shape
• Vision
• Sensitivity to color
• Adjustable field of view
• Movement / Turning ability
• Eating / Mating / Attacking / Lighting
• Energy provides life
• 2 types of energy stored and ready

8
Critter Design - Mental

• BCM like neural network brain
• Model developed to approximate neurons in the
  visual cortex
• Adapt to changing inputs plasticity
• Vision and Energy inputs
• Move / Eat / Attack, etc. outputs
• Neurons appear in groups
• 10 32 neuron groups and same for neurons in
  groups
• Neurons excitatory or inhibitory

9
(No Transcript)
10
Critter Design - Evolution

• Employs standard genetic algorithm
• No explicit fitness function
• Fitness evaluated by passing along your genes
• Crossover / Mutation of genes
• Critter described by its genome
• 1460 genes
• Describe all physical / mental aspects

11
Critter Design - Evolution

• Physical genes
• Energy usage rates
• Base metabolism / Max energy usage
• Indirectly describe size / strength
• Mental genes
• Describe general layout of brain and its
  interconnections
• Brain grown from these parameters no two
  alike

12
Architectural Design

• Distributed system with multiple cross-platform
  clients (Windows / Linux / Solaris)
• Server handles rendering the world and
  interactions
• Clients process the neural networks
• Real-time analysis client
• IPC network protocol
• Library by Reid Simmons (CMU/RI)
• OpenGL rendering (5 15 frames/sec.)
• User display and each critters view
• Movie output (AVI format)

13
Analysis

• Dumping of individual brains in multiple formats
• Plaintext (in the future import brains)
• HTML (group connectivity overview)
• .GDL (graphical layout)
• Dumping of critter genome
• Real-time dumping of various system-wide
  statistics
• HTML with JPEGs
• Num. births / deaths, avg. critter energy, etc.

14
Analysis (cont)

• Movie output
• Speeds up visual observation
• Keeps record of interesting behavior
• Critter selection / observation
• Behind-the-shoulder view
• Eye view
• Various statistics

15
Lasers

• Greg got bored and made our simulator a game
• You were the only one to have a weapon
• It was a laser
• It was red
• It killed the other critters
• Playtesting currently in progress

16
Behaviors

• Interesting to note tendency of critters to
  always be turning
• Caused by the way the turn behavior is expressed
• Observed behaviors
• Grazing critter slows down when near food, eats
  multiple observations
• Prolific mating

17
Future Work

• Getting all the bugs out
• More analysis tools
• Cross-generation genome analysis
• Longer test-runs
• Testing fitness
• Placing existing critter in new environment
• Mixing separately-evolved populations
• Increased performance

18
(No Transcript)