Title: CAP6938 Neuroevolution and Artificial Embryogeny Competitive Coevolution
1CAP6938Neuroevolution and Artificial
EmbryogenyCompetitive Coevolution
- Dr. Kenneth Stanley
- February 20, 2006
2ExampleI Want to Evolve a Go Player
- Go is one of the hardest games for computers
- I am terrible at it
- There are no good Go programs either
(hypothetically) - I have no idea how to measure the fitness of a Go
player - How can I make evolution solve this problem?
3Generally Fitness May Be Difficult to Formalize
- Optimal policy in competitive domains unknown
- Only winner and loser can be easily determined
- What can be done?
4Competitive Coevolution
- Coevolution No absolute fitness function
- Fitness depends on direct comparisons with other
evolving agents - Hope to discover solutions beyond the ability of
fitness to describe - Competition should lead to an escalating arms race
5The Arms Race
6The Arms Race is an AI Dream
- Computer plays itself and becomes champion
- No need for human knowledge whatsoever
- In practice, progress eventually stagnates
(Darwen 1996 Floreano and Nolfi 1997 Rosin and
Belew 1997)
7So Who Plays Against Whom?
- If evaluation is expensive, everyone cant play
everyone - Even if they could, a lot of candidates might be
very poor - If not everyone, who then is chosen as
competition for each candidate? - Need some kind of intelligent sampling
8Challenges with Choosing the Right Opponents
- Red Queen Effect Running in Circles
- A dominates B
- C dominates B
- A dominates B
- Overspecialization
- Optimizing a single skill to the neglect of all
others - Likely to happen without diverse opponents in
sample - Several other failure dynamics
9Heuristic in NEATUtilize Species Champions
Each individual plays all the species champions
and keeps a score
10Hall of Fame (HOF)(Rosin and Belew 1997)
- Keep around a list of past champions
- Add them to the mix of opponents
- If HOF gets too big, sample from it
11More RecentlyPareto Coevolution
- Separate learners and tests
- The tests are rewarded for distinguishing
learners from each other - The learners are ranked in Pareto layers
- Each test is an objective
- If X wins against a superset of tests that Y wins
again, then X Pareto-dominates Y - The first layer is a nondominated front
- Think of tests as objectives in a multiobjective
optimization problem - Potentially costly All learners play all tests
De Jong, E.D. and J.B. Pollack (2004). Ideal
Evaluation from Coevolution Evolutionary
Computation, Vol. 12, Issue 2, pp. 159-192,
published by The MIT Press.
12Choosing Opponents Isnt Everything
- How can new solutions be continually created that
maintain existing capabilities? - Mutations that lead to innovations could
simultaneously lead to losses - What kind of process ensures elaboration over
alteration?
13Alteration vs. Elaboration
14Answer Complexification
- Fixed-length genomes limit progress
- Dominant strategies that utilize the entire
genome must alter and thereby sacrifice prior
functionality - If new genes can be added, dominant strategies
can be elaborated, maintaining existing
capabilities
15Test Domain Robot Duel
- Robot with higher energy wins by colliding with
opponent - Moving costs energy
- Collecting food replenishes energy
- Complex task When to forage/save energy,
avoid/pursue?
16Robot Neural Networks
17Experimental Setup
- 13 complexifying runs, 15 fixed-topology runs
- 500 generations per run
- 2-population coevolution with hall of fame (Rosin
Belew 1997)
18Performance is Difficult to Evaluate in
Coevolution
- How can you tell if things are improving when
everything is relative? - Number of wins is relative to each generation
- No absolute measure is available
- No benchmark is comprehensive
19Expensive Method Master Tournament(Cliff and
Miller 1995 Floreano and Nolfi 1997)
- Compare all generation champions to each other
- Requires n2 evaluations
- An accurate evaluation may involve e.g. 288 games
- Defeating more champions does not establish
superiority
20Strict and Efficient Performance Measure
Dominance Tournament (Stanley Miikkulainen
2002)
21Result Evolution of Complexity
- As dominance increases so does complexity on
average - Networks with strictly superior strategies are
more complex
22Comparing Performance
23Summary of Performance Comparisons
24The Superchamp
25Cooperative Coevolution
- Groups attempt to work with each other instead of
against each other - But sometimes its not clear whats cooperation
and whats competition - Maybe competitive/cooperative is not the best
distinction? - Newer idea Compositional vs. test-based
26Summary
- Picking best opponents
- Maintaining and elaborating on strategies
- Measuring performance
- Different types of coevolution
- Advanced papers on coevolution
Ideal Evaluation from Coevolution by De Jong,
E.D. and J.B. Pollack (2004)Monotonic Solution
Concepts in Coevolution by Ficici, Sevan G.
(2005)
27Next Topic Real-time NEAT (rtNEAT)
- Simultaneous and asynchronous evaluation
- Non-generational
- Useful in video games and simulations
- NERO Video game with rtNEAT
-Shorter symposium paper Evolving Neural Network
Agents in the NERO Video Game by Kenneth O.
Stanley and Risto Miikkulainen (2005)-Optional
journal (longer, more detailed) paper Real-time
Neuroevolution in the NERO Video Game by Kenneth
O. Stanley and Risto Miikkulainen (2005)
-http//Nerogame.org -Extra coevolution papers
Homework due 2/27/06 Working genotype to
phenotype mapping. Genetic representation
completed. Saving and loading of genome file I/O
functions completed. Turn in summary, code, and
examples demonstrating that it works.
28Project Milestones (25 of grade)
- 2/6 Initial proposal and project description
- 2/15 Domain and phenotype code and examples
- 2/27 Genes and Genotype to Phenotype mapping
- 3/8 Genetic operators all working
- 3/27 Population level and main loop working
- 4/10 Final project and presentation due (75 of
grade)