fmrg-ballet's Introduction

Bayesian Optimization with Adaptive Level-Set Estimation (BALLET)

Project for Protein study. README written in Markdown. The PDF file for the recent submission to ICML ReALML workshop is attached in the folder.

Environment

The implementation has been verfied on python 3.9.7 MacOS 12.4 (M1 Pro). The required packages are included in requirements.txt, to install all required packages, please run the following code.

pip install -r requirements.txt

Guidance

One example is shown below. The algorithm will extract the data from the directory specified by datadir and return the suggested candidate to evaluate.

python main.py --name="exp_name" --aedir="./tmp/tmp_ae" --subdir="./res/"  --datadir="./data/Capacity_example.xlsx" --batch-size=5 --train_times=10 --beta=10 --acq_func="ucb" --learning_rate=6 -f -a;

# example of corresponding output:
       Biomass source  Highest temp. (oC)  Fe load (mC%)  KOH load (mC%)
0                MxG               800.0            6.0            60.0
1  Commercial Lignin              1100.0            1.0            40.0
2               Hemp              1000.0           12.0           100.0
3               Hemp              1200.0            6.0            40.0
4  Commercial Lignin              1100.0            3.0            40.0

Note: valid value of each attributes (2160 in toal)

_biomass_categories     = ['Switchgrass', 'Hemp', 'MxG', 'Commercial Lignin']
_temp                   = [600, 700, 800, 900, 1000, 1100, 1200, 1300, 1400, 1500]
_fe                     = [200, 100, 50, 25, 12, 6, 3, 2, 1]
_koh                    = [0, 20, 40, 60, 80, 100]

Recommend Projects