Advisor: Dr. Mirzaei MohammadTaghi Moein Isfahan University of Technology
The Scheme of Evolution Model Current Population New Individuals New Population
Generating New Individuals in Darwinian-Type Evolution Model Current Population Generate New Individuals by mutation and recombination Candidate Parents
Generating New Individuals in LEM Current Population L-group Low Performance Individuals H-group High Performance Individuals Generate hypothesis for H-group Generate hypothesis for L-group Generate New Individuals by Instantiating the hypothesis
Extrema Generation fitness-based according to two thresholds, called HFT and LFT
Extrema Generation Cont. population-based according to two parameters, called HPT and LPT
Extrema Generation Cont. The fitness-based and population-based methods can also be used in combination. a global approach applies one of the above methods to the entire population. a local approach applies one of the above methods in parallel to different subsets of the population. The above methods can be enhanced by employing elitism.
Extrema Generation Cont. In the above methods, the H-group and L-group were selected only from the current population. H-group description that does not take into consideration past L-groups is likely to be too general. L-group description that does not take into consideration past L-groups is likely to be too specific.
Considering History of Evolution Population-lookback union of the past L-groups plus the L-group in the current population is the actual L-group. The number of past L-groups is specified by the p-lookback parameter. High-group description-lookback current H-group description is used to generate new candidate individuals. past H-group descriptions serve as preconditions for accepting a candidate. The number of H-group descriptions is specified by the d- lookback parameter.
Considering History of Evolution Cont. Low-group description-lookback maintains a collection of descriptions of L-groups. uses them as constraints when generating H-group descriptions. Incremental specialization uses incremental learning algorithm to maintain one updated description of the H-group. input to such an algorithm is a description of the previous H-group.
Generating Description(AQ) Seed Selection Star Generation Rule Selection Coverage Update Finish No Yes any positive example
Description instantiation New individuals should satisfy all H-group descriptions. A description instantiation is done by assigning different combinations of values to variables in the rules of a ruleset. Each assignment must satisfy all conditions in at least one of the rules.
LEM Algorithm 1. Generate a population 2. Execute machine learning mode a) Derive extrema b) Create a hypothsis c) Generate new individuals d) Go to step (2-a) and continue until termination condition is met, if termination condition is met do: i. If the LEM termination condition is met, end the evolution. ii. Repeate the process from step 1, this is called start-over. iii. Go to step 3
LEM Algorithm Cont. 3. Execute Darvinian Evolution mode 4. Alternate: Go to step 2, and continue alternating between step 2 and step 3 until the LEM termination condition is met.
Generating start-over population A. Select-elite B. Avoid-past-failures C. Use-recommendatoins D. Generate-a-variant
Any Questions? Thanks for your attention