JqVicky/actlearn_optint

Active learning for optimal intervention design in causal models

Active Learning for Optimal Interventions

Code for paper: Active learning for optimal intervention design in causal models

arXiv link: coming soon...

Installation

Download code and run in command line:

pip install -e .

Run on a synthetic instance, e.g.:

python run.py --nnodes 5 --noise_level 1 --DAG_type path --std --a_size 2 --a_target 3 4 --acquisition greedy

Source code folder: ./optint/

More examples given in: ./optint/notebook/test_multigraphs.ipynb

Examples on Perturb-CITE-Seq [1]

Source code folder: ./perturb-CITE-seq

Notebooks for exploratory data analysis: ./perturb-CITE-seq/preprocess

• download and extract data: ./perturb-CITE-seq/preprocess/screen_sanity_checks.ipynb
• process data: ./perturb-CITE-seq/preprocess/process_data.ipynb

Notebook for running the optimal intervention design task: ./perturb-CITE-seq/test.ipynb

Figures in the paper

Illustraive figures: made using mac keynotes

Pointers for nonillustrative figures:

• ./optint/notebook/test_ow.ipynb: Fig. 3, S2
• ./optint/notebook/test_convergence.ipynb: Fig. 4
• ./optint/notebook/test_multigraphs.ipynb: Fig. 5, S4-7
• ./perturb-CITE-seq/preprocess/screen_sanity_checks.ipynb: Fig. S8, S10, S11A
• ./perturb-CITE-seq/preprocess/process_data.ipynb: Fig. S9
• ./perturb-CITE-seq/preprocess/test_linearity.ipynb: Fig. S11C
• ./perturb-CITE-seq/test.ipynb: Fig. 6, S12-15