230 likes | 349 Views
Interactive Evolutionary Computation. As we have seen in this course, EC is a powerful search paradigm. As long as the user is able to develop an adequate evaluation function, the ECs studied so far will perform well.
E N D
Interactive Evolutionary Computation • As we have seen in this course, EC is a powerful search paradigm. • As long as the user is able to develop an adequate evaluation function, the ECs studied so far will perform well. • However, what happens when on cannot develop an evaluation function that is an accurate closed-form mathematical equation?
Interactive Evolutionary Computation • Interactive Evolutionary Computation (IEC) is a welcomed alternative when: • An accurate evaluation function is difficult to develop, and • The user has an idea of what a good solution may look like.
Interactive Evolutionary Computation • IECs have been used for wide variety of applications including [Takagi, H. (2001). “Interactive Evolutionary Computaton: Fusion of the Capabilities of EC Optimization and Human Evaluation”, Proceedings of the IEEE, pp. 1275-1296, Vol. 89, No. 9, September, IEEE Press]: • Graphic Art, Computer Graphics, Animation, • Music, • Editorial Design, • Industrial Design, • Face Image Generation, • Speech Processing, • Hearing Aid Adaptation, • Database Retrieval, • Data Mining • Image Processing, • Robotics, • etc
Interactive Evolutionary Computation • Not all IECs are the same. There seems to be a continuum. Kennedy, Externalized Particle Swarm Takagi, Lund, etc Parmee EC Intensive Human Interaction Intensive Human/EC Collaborative
Interactive Evolutionary Computation • IECs are not just limited to human evaluation. • Human evaluation functions are used to provide a fitness for individuals being evolved. • Instead using human evaluation, human selection (algorithms) can be used. • Also, human guided procreation has been used in IECs as well.
Interactive Evolutionary Computation Examples • Brian Carnahan’s IGA for Anthropomorphic Design (Overcome by Fumes)
Interactive Distributed Evolutionary Algorithms (IDEAs) • Interactive Evolutionary Computation has been shown to be a very powerful technique for solving design problems where the fitness function cannot be expressed as a closed form mathematical equation. • Research in the area of Distributed and Parallel Evolutionary Computation has been successfully used to speed up the evolutionary search.
Interactive Distributed Evolutionary Algorithms (IDEAs) • Therefore, interactive distributed evolutionary computation (IDEC) holds a great deal of promise because: • Multiple users should be able to design artifacts more quickly than a single user (reducing user fatigue), • Artifacts developed by multiple users will have a wider range of acceptance, • By observing how humans interactively solve problems, we may gain a better understanding (heuristics?) of how to develop ‘intelligent’ ECs.
An IDEA for Emoticon Design Procedure IDEA_Client{ t = 0; Initialize Pop(t) // Randomly Generate 9 Emoticons; Present Pop(t) to User; While (Not Done) { Allow_User_to_Select_An_Emoticon(e); Allow_User_to_Select_A_Mutation_Op(o); Send_to_Meme_Space(e); Receive_From_Meme_Space(m); Parents(t) = {e, m}; Offspring(t) = {Create_4_Mutants(e,o); Create_3_Recombinants(e,m,o);} Pop(t+1) = Parents(t) Offspring(t); t = t + 1; } }
An IDEA for Emoticon Design • http://iis.cse.eng.auburn.edu/~gvdozier/IDEA-1.html
An IDEA for Emoticon Design • Representation of a Candidate Emoticon
Experiments • For a proof of concept we conduct 3 simple experiments (where Meme Space = 2.5 x # of Networked Users) : • Smiley Face • Anger • Hand-In-Gear
Results: Smiley Face • Which is the Best Smiley Face?
Results: Anger • Which of these is the Best Anger Emoticon?
Results: Hand-In-Gear • Which of these is the Best Hand-In-Gear Emoticon
Conclusions • Our results show that IDEA can be used to allow multiple users to interactively design emoticons. • Using C4.5, the emoticons can be separated into two groups. The differences between the two groups are statistically significant.
Discussion: Why Use an IDEA ? • Not All Designers are the Same • Starters • Middle Relievers • Relievers (Finishers) • Not All Designers Work at the Same Pace • Not All Designers Have the Same Gifts • Some designers may be better critics (selectors) • Some designers may be more skilled with the evolutionary operators • Exploits Design Team Diversity • Meme Space: • Allows a form of Evolutionary Backtracking • Allows more universal ideas to survive longer
IDEA Questions • Some questions concerning the use of IDEAs are as follows: • Given a design team of N users what is the most effective meme space size? • How do we make sure that faster users don’t overrun the meme space? • What is the best composition for a design team?
Some Questions (cont.) • How many designs should one receive from meme space on a given generation? • How does this number change over the evolutionary process? • Should meme space designs be crossed with user selected images?
Some Questions (cont.) • Should the amount of crossover change over the evolutionary process? • How can we detect and reduce user fatigue and frustration?