An api and a collection physics-based uncertainty-aware models to drive automated CO2RR laboratories
See our paper: A physics-based data-driven model for CO2 gas diffusion electrodes to drive automated laboratories
pip install git+https://github.com/igrega348/CO2-catalysis.git@active_learning_api
from carbondriver import GDEOptimizer
from carbondriver.loaders import load_data
gde = GDEOptimizer("MLP", output_dir="./tmp_test_out")
_, _, _, _, df = load_data("paper/Characterization_data.xlsx")
df_train = df.loc[:30]
df_explore = df.loc[31:]
ei, next_pick = gde.step_within_data(df_train, df_explore)
print("The next experiment to try is the following:")
print(df_explore.iloc[int(next_pick)])