This repository contains the code used for the experiments in "Estimating alpha-Rank by Maximising Information Gain". Paper is available here.
In particular it contains:
- An implementation of ResponseGraphUCB.
- A parallelized implementation of \alphaIG and \alphaWass.
- The X Good, Y Bad payoff games used in the paper.
To run experiments please use the run_experiments.py file.
All hyper-parameters and settings are specified in run_experiments.py.
In order to change the game used, please adjust the env_params dictionary (the settings for 3 Good, 5 Bad are currently specified with the settings for 2 Good, 2 Bad and 4x4 Gaussian game being commented out).
The exp_params list contains a list of algorithm configs (specified via dictionary) that will all be run.
If a list is specified as an argument, then all values in that list will be run (as is the case for \delta in ResponseGraphUCB).
The default hyper-parameters used for the 3 Good, 5 Bad experiments are currently used for the algorithms. Please change appropriately.
The notebook notebooks/graphing.ipynb contains the code used for generating the regret graphs in the paper.
run_experiments.py saves pickled dictionaries that contain all the information generated by a run.
Please see sampling.py for the code that actually generates the data, and the exact meaning of each field.
All Python3 packages used are specified in requirements.txt.
You will likely need install numpy before attempting to install ndd.
Additionally, a fortran compiler is required for ndd.
This can be installed on Ubuntu via gfortran (apt install gfortran).
This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.
When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.
This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact opencode@microsoft.com with any additional questions or comments.
This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow Microsoft's Trademark & Brand Guidelines. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos are subject to those third-party's policies.