Martijho-PathNet-search-plots

(Forskjeller mellom versjoner)
Gå til: navigasjon, søk
 Versjonen fra 28. feb 2018 kl. 16:03 (vis kilde) (→Diversity metrics)← Eldre redigering Nåværende revisjon fra 28. feb 2018 kl. 16:12 (vis kilde) Linje 1: Linje 1: - == Population Diversity == + = Population Diversity = - === Diversity metrics === + == Diversity metrics == Three different diversity calculations have been performed, which resulted in three similar diversity messures: Three different diversity calculations have been performed, which resulted in three similar diversity messures: {|style="margin: 0 auto;" {|style="margin: 0 auto;" Linje 8: Linje 8: |} |} - === Home-made for each layer === + == Home-made for each layer == {|style="margin: 0 auto;" {|style="margin: 0 auto;" | [[File:Population_diversity_recomb.png|300px|thumb|upright|alt=Population diversity in recomb search|Average population diversity for each layer in generation. Experiment: Recomb]] | [[File:Population_diversity_recomb.png|300px|thumb|upright|alt=Population diversity in recomb search|Average population diversity for each layer in generation. Experiment: Recomb]] Linje 19: Linje 19: |} |} - == Path size == + = Path size = {|style="margin: 0 auto;" {|style="margin: 0 auto;" | [[File:Average_path_size.png|300px|thumb|upright|alt=Path size|The change in average path size for each generation. ]] | [[File:Average_path_size.png|300px|thumb|upright|alt=Path size|The change in average path size for each generation. ]] Linje 25: Linje 25: |} |} - == Capacity and Reuse == + = Capacity and Reuse = {|style="margin: 0 auto;" {|style="margin: 0 auto;" | [[File:Capacity_pr_generation.png|300px|thumb|upright|alt=Capacity development during search|Average number of modules used during the different searches plotted consecutively. The gray dotted line is what number of modules would be used if modules were selected randomly for each task.]] | [[File:Capacity_pr_generation.png|300px|thumb|upright|alt=Capacity development during search|Average number of modules used during the different searches plotted consecutively. The gray dotted line is what number of modules would be used if modules were selected randomly for each task.]] Linje 32: Linje 32: |} |} - == Training == + = Training = {|style="margin: 0 auto;" {|style="margin: 0 auto;" | [[File:Training_accuracy.png|300px|thumb|upright|alt=Training accuracy|Average training accuracy for each of the five search algorithms during the search for each task. The dotted line is the validation accuracy reached after the search was completed.]] | [[File:Training_accuracy.png|300px|thumb|upright|alt=Training accuracy|Average training accuracy for each of the five search algorithms during the search for each task. The dotted line is the validation accuracy reached after the search was completed.]]

Population Diversity

Diversity metrics

Three different diversity calculations have been performed, which resulted in three similar diversity messures:

 Pairwise Hamming distance: Calculated for each layer and then averaged across each path. This plot took 22 minutes and 40 seconds to generate. Homemade diversity metric: This diversity metric resulted in a scaled version of Pairwise Hamming, but took only 36 seconds to generate. See description below Average euclidean distance to centroid: Calculated by averaging the euclidean distance between every genotype in a generation, and a "centroid genotype". The plot was generated in 40 seconds

 Average population diversity for each layer in generation. Experiment: Recomb Average population diversity for each layer in generation. Experiment: Low Average population diversity for each layer in generation. Experiment: High
 Average population diversity for each layer in generation. Experiment: Low to High Average population diversity for each layer in generation. Experiment: High to Low

Path size

 The change in average path size for each generation. The average size of each layer in each optimal path found.

Capacity and Reuse

 Average number of modules used during the different searches plotted consecutively. The gray dotted line is what number of modules would be used if modules were selected randomly for each task. Used capacity in a PathNet plotted against that networks average classification accuracy for all tasks Average reuse for each generation plotted for each experiment type. The gray dotted line is what reuse would be reached by random module selection

Training

 Average training accuracy for each of the five search algorithms during the search for each task. The dotted line is the validation accuracy reached after the search was completed. Number of training iterations in total for all modules in the PathNet plotted against the cumulative validation accuracy reached for all tasks. The size of the circle corresponds to the amount of modules used. Plot of the ratio of unit training in locked and stored modules on the total number of unit training in all modules for each task