https://github.com/giuse/dne
A set of neuroevolution experiments with/towards deep networks
https://github.com/giuse/dne
rubyml
Last synced: 3 months ago
JSON representation
A set of neuroevolution experiments with/towards deep networks
- Host: GitHub
- URL: https://github.com/giuse/dne
- Owner: giuse
- License: mit
- Created: 2018-03-05T13:52:40.000Z (almost 8 years ago)
- Default Branch: master
- Last Pushed: 2019-12-31T12:47:23.000Z (about 6 years ago)
- Last Synced: 2024-01-23T09:59:50.166Z (almost 2 years ago)
- Topics: rubyml
- Language: Ruby
- Size: 48.8 KB
- Stars: 125
- Watchers: 5
- Forks: 10
- Open Issues: 3
-
Metadata Files:
- Readme: README.md
- License: LICENSE.txt
Awesome Lists containing this project
README
# Deep Neuroevolution experiments
This project collects a set of neuroevolution experiments with/towards deep networks for reinforcement learning control problems using an unsupervised learning feature exctactor.
## *Playing Atari with Six Neurons*
The experiments for this paper are based on [this code](https://github.com/giuse/DNE/releases/tag/six_neurons).
The algorithms themselves are coded in the [`machine_learning_workbench` library](https://github.com/giuse/machine_learning_workbench), specifically using [version 0.8.0](https://github.com/giuse/machine_learning_workbench/releases/tag/0.8.0).
## Installation
First make sure the OpenAI Gym is pip-installed on python3, [instructions here](https://github.com/openai/gym).
You will also need the [GVGAI_GYM](https://github.com/rubenrtorrado/GVGAI_GYM) to access GVGAI environments.
Clone this repository, then execute:
$ bundle install
## Usage
bundle exec ruby experiments/cartpole.rb
## Contributing
Bug reports and pull requests are welcome on GitHub at https://github.com/giuse/DNE.
## License
The gem is available as open source under the terms of the [MIT License](https://opensource.org/licenses/MIT).
## References
Please feel free to contribute to this list (see `Contributing` above).
- **UL-ELR** stands for Unsupervised Learning plus Evolutionary Reinforcement Learning, from the paper _"Intrinsically Motivated Neuroevolution for Vision-Based Reinforcement Learning" (ICDL2011)_. Check [here](https://exascale.info/members/giuseppe-cuccu/) for citation reference and pdf.
- **BD-NES** stands for Block Diagonal Natural Evolution Strategy, from the homonymous paper _"Block Diagonal Natural Evolution Strategies" (PPSN2012)_. Check [here](https://exascale.info/members/giuseppe-cuccu/) for citation reference and pdf.
- **RNES** stands for Radial Natural Evolution Strategy, from the paper _"Novelty-Based Restarts for Evolution Strategies" (CEC2011)_. Check [here](https://exascale.info/members/giuseppe-cuccu/) for citation reference and pdf.
- **Online VQ** stands for Online Vector Quantization, from the paper _"Intrinsically Motivated Neuroevolution for Vision-Based Reinforcement Learning" (ICDL2011)_. Check [here](https://exascale.info/members/giuseppe-cuccu/) for citation reference and pdf.
- The **OpenAI Gym** is described [here](https://gym.openai.com/) and available on [this repo](https://github.com/openai/gym/)
- **PyCall.rb** is available on [this repo](https://github.com/mrkn/pycall.rb/).