/errp_decode

Codes and data analysis for workshop wearsys 2020: "Human-In-The-Loop RL with an EEG Wearable Headset: On Effective Use of Brainwaves to Accelerate Learning"

Primary LanguageJupyter Notebook

wearsys20

Codes and data analysis for workshop wearsys 2020: "Human-In-The-Loop RL with an EEG Wearable Headset: On Effective Use of Brainwaves to Accelerate Learning"

Running codes

Put codes in train/ folder, with a separate folder for each subject

  • Run errp_riemann.ipynb for running the baseline (state-of-the-art) results
  • Run errp_proposed.ipynb for running the proposed algorithm

Note:

  1. Choose flag_test=False for running 10-fold CV results (Paper includes only 10-fold CV)

Data Analysis

All the results and analysis is in data_analysis/ folder

Motivation Results

  • Activate conda environment ml-py37-gpu and run rl_errp.py
  • Change p_err_sim flag to the decoding accuracy of errp (e.g., 0.6,0.7, etc.)
  • Copy/paste the values to plot folder, and run plot_erp_arl_all.py