[mlpack-git] [mlpack/mlpack] NEAT algorithm (#752)

Excalibur notifications at github.com
Fri Aug 5 16:01:00 EDT 2016


Implemented NEAT algorithm, and multiple test cases.
You can view, comment on, or merge this pull request online at:

  https://github.com/mlpack/mlpack/pull/752

-- Commit Summary --

  * change .gitignore to ignore .DS_Store
  * create gene.hpp, add ne_test.cpp, and created or modified multiple CMakeLists.txt.
  * Created gene.happ, add tests and changed or created CMakeLists.txt
  * Add aDepth to NeuronGene.
  * implement CNE, haven't finish.
  * delete blank files
  * finished genome.hpp, revised gene.hpp
  * modified gene.hpp
  * implemented CNE skeleton; more genome and population functions.
  * implemented XOR test case; need debugging; one heart function -- CNE::Evolve() remains to fulfill.
  * solved most bugs for testing XOR
  * further solved some bugs in code.
  * Solved bugs. Need to fulfill Reproduce() to finish CNE.
  * CNE algorithm finished. XOR test passes. Need to revise coding style and some TODOs. Then continue other algorithms.
  * Create Species class.
  * Revised LinkGene and related file for NEAT
  * revised sortSpecies()
  * add aAdjustedFitness to genome
  * fix bug in CNE Reproduce()
  * Implemented neat mutations, replace size_t by ssize_t
  * Implemented crossover, some bug exist
  * Revise Crossover, solve bugs, more functions implementations.
  * revised neat functions. In paper,disabled links also consider for crossover and calc distance.
  * neat almost finished. Need debugging,need revise coding style, need testing.
  * NEAT finished, but has bugs. XOR not pass.
  * In progress of debugging. Add lots of printf ... Will clear them after debugging
  * seems bugs solved. keep printf. Seems information such as species size, numSpecies are not updated when things changed, that is why so many bugs.
  * solve more bugs
  * solved more. But still have problem.
  * solved CalcAverageRank bug. Still have more.
  * currently seems no bug. Previously we removed stale species even when species number is small, leads to species disappear.
  * revised WeightDiff
  * Removed redundant members in species, population. haven't remove in genome, as it caused some bug. Still need checking.
  * Further cleaned some code.
  * changed the place where we should set the childGenome's NumInput, NumOutput, GenomeDepth. However, seems calculate genome depth is quite slow.
  * changed neuron_gene, genome's activation. Still need to revise neat.
  * Revised MutateAddLink.
  * solved some bugs about activation.
  * Add Cart Pole test. Result not good. Fitness doesn't improve.
  * Solved more bugs.
  * Fixed a bug in task. Removed adjustFitness. Problems maybe the parameter setting.
  * further revised something.
  * Passed Cart Pole problem. Revised neat, task.
  * Implemented test MountainCar, not pass yet.
  * Seems passed Mountain Car.. in 1 iteration
  * Revised a bug in TaskMountainCar Action()
  * added Success() for tasks.
  * Finished double pole. Non-Markov have bug.
  * revised non-markov, still have bug
  * revised neat MutateAddLink. Non-Markov still have bug.
  * revised neat, clean genome before evaluate.
  * Merge pull request #1 from mlpack/master
  * revised a bug in Activate()
  * fix conflict
  * changed header file in ne_test

-- File Changes --

    M .gitignore (2)
    M src/mlpack/methods/CMakeLists.txt (1)
    A src/mlpack/methods/ne/CMakeLists.txt (23)
    A src/mlpack/methods/ne/cne.hpp (186)
    A src/mlpack/methods/ne/genome.hpp (334)
    A src/mlpack/methods/ne/link_gene.hpp (111)
    A src/mlpack/methods/ne/neat.hpp (976)
    A src/mlpack/methods/ne/neuron_gene.hpp (174)
    A src/mlpack/methods/ne/parameters.hpp (96)
    A src/mlpack/methods/ne/population.hpp (154)
    A src/mlpack/methods/ne/species.hpp (141)
    A src/mlpack/methods/ne/tasks.hpp (692)
    A src/mlpack/methods/ne/utils.hpp (88)
    M src/mlpack/tests/CMakeLists.txt (5)
    A src/mlpack/tests/ne_test.cpp (691)

-- Patch Links --

https://github.com/mlpack/mlpack/pull/752.patch
https://github.com/mlpack/mlpack/pull/752.diff

---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/mlpack/mlpack/pull/752
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160805/00b3833d/attachment-0001.html>


More information about the mlpack-git mailing list